<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
727622
Z Image
Ultra-realistic 8K scene of the person in the photo, stepping out of and breaking the panel border of a giant open manga book lying on the floor of an aesthetic bedroom. The comic pages depict a surreal 4-panel black-and-white manga with screentones, bold ink lines, and authentic Japanese layout, featuring the manga version of the same person trapped inside. The larger, real-life version emerges dynamically into the comic’s surface, vibrant, colorful, and photorealistic, in sharp contrast to the monochrome manga world. Both versions interact directly, creating a meta, humorous encounter. Highly detailed, crisp manga line art seamlessly integrated with vivid, realistic
432772
Nano Banana
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
507522
Nano Banana
Ultra-realistic 8K scene of the person in the photo, stepping out of and breaking the panel border of a giant open manga book lying on the floor of an aesthetic bedroom. The comic pages depict a surreal 4-panel black-and-white manga with screentones, bold ink lines, and authentic Japanese layout, featuring the manga version of the same person trapped inside. The larger, real-life version emerges dynamically into the comic’s surface, vibrant, colorful, and photorealistic, in sharp contrast to the monochrome manga world. Both versions interact directly, creating a meta, humorous encounter. Highly detailed, crisp manga line art seamlessly integrated with vivid, realistic
24964
Nano Banana
Steps: 20, Sampler: DPM++ 2M Karras, CFG scale: 7, Seed: 413127992, Size: 1280x1600, Model hash: 7fa552ffb0, Model: nijidiffusedmix_v40, Denoising strength: 0.3, Clip skip: 2, ADetailer model: face_yolov8s.pt, ADetailer prompt: "dark blue eyes, freckles below eyes", ADetailer confidence: 0.3, ADetailer dilate erode: 4, ADetailer mask blur: 4, ADetailer denoising strength: 0.5, ADetailer inpaint only masked: True, ADetailer inpaint padding: 32, ADetailer version: 23.12.0, Version: v1.6.1
10493
Stable Diffusion
gett nude Learn to create better images with AI! Check out our crash course in prompt engineering & AI art generation! AM1970 posted about 2 hours ago 8 views 0 comments Behold this realistic masterpiece: a slim, elegant woman with fiery red hair tied in a long ponytail, defined eyebrows framing her striking blue eyes, and pale skin dotted with charming freckles across her face and chest. Captured from a side angle, she radiates happiness with a closed-mouth smile, her upturned nose adding to her allure. Dressed in a vibrant, colorful fabric robe pieced together with zips, buttons, seams, and intricate stitching—fully clothed yet teasing partial nudity with subtle cleavage and teardrop-shaped, saggy breasts for a natural, confident vibe. Illuminated by dramatic rim lighting that highlights every textured, detailed inch of her perfect skin, this high-quality image is a celebration of beauty, realism, and artistry! Read more... Comments (0) Add a comment Stunning handsome readhead woman by AM1970 • Follow 0 realistic, masterpiece, high quality BREAK 1woman, 28 y. o. slim fit woman, red hair, long pony tail, defined eyebrows, (fully clothed) , pale skin, (realistic skin:1.3) | (textured skin:1.3) | (detailed skin:1.3) | (perfect skin:1.3) , (upturned nose) , freckles on face and chest, colorful fabric combining to form robe zips, buttons, seams, stitching), partial nudity, teardrop boobs, blue eyes, (cleavage:1.2) , saggy breasts. side angle. happy| closed mouth. ((rim lighting) ) Negative prompt: unrealistic, saturated, high contrast, big nose, painting, drawing, sketch, cartoon, anime, manga, render, CG, 3d, watermark, signature, label, nsfw, nudity, fetish, latex, pvc, suggestive pose, weapon, combat, blood, war scene, aggressive expression, harsh shadows, heavy makeup, color tint, saturation, chromatic aberration, lens flare, low resolution, blur, noise, artifacts, over-retouching, plastic skin, oversharpening, watermark, text, logo Generation parameters 832x1216 Model used CyberRealistic XL v7.0 [Download AI model] More by AM1970 Twitter Instagram Say hi 👋 PromptHero is built by ElevenYellow . Navigation Create AI images Academy - Learn AI PromptHero Community Affiliate program Openjourney AI Challenges Verified profiles Plans Advertise FAQ 🟢 Status Learn AI Content Creation with AI for Brands and Products Midjourney for graphic design & art professionals Crash course in generative AI & prompt engineering for images AI influencers and consistent characters Create custom AI models and LoRas by fine-tuning Stable Diffusion Master your composition: advanced AI image generation with ControlNet Guides Best AI Upscalers 2025 Best AI Video Editors 2025 Best AI 3D Models 2025 Best AI Image Editors 2025 Best AI Notetakers 2025 The Best AI Prompts Best AI Prompts Best Veo Prompts Best Hailuo Prompts Best Wanx Prompts Best Hunyuan Prompts Best Seedance Prompts Best Pixverse Prompts Best Luma Prompts Best Kling Prompts Best Runway Prompts Best Nano Banana Prompts Best SeedEdit Prompts Best Qwen Prompts Best FLUX Prompts Best Recraft Prompts Best Ideogram Prompts Best Seedream 4 Prompts Best Google Imagen Prompts Best ChatGPT Image Prompts Best Stable Diffusion Prompts Best DreamShaper Prompts Best Realistic Vision Prompts Best epiCRealism Prompts Best majicMIX realistic Prompts Best Midjourney Prompts Best Openjourney Prompts Best ChilloutMix Prompts Best Deliberate Prompts Best ReV Animated Prompts Best Perfect World Prompts Best Dreamlike Photoreal Prompts Best Dreamlike Diffusion Prompts Best Anything Prompts Best Protogen Prompts Best AbyssOrangeMix Prompts Best Grapefruit Prompts Best Kenshi Prompts Best Analog Diffusion Prompts Best SynthwavePunk Prompts Best Illuminati Diffusion Prompts Best Arcane Diffusion Prompts Best Mo-Di Diffusion Prompts Best Redshift Diffusion Prompts Best Waifu Diffusion Prompts Best TrinArt Prompts Best Poolsuite Diffusion Prompts Best Funko Diffusion Prompts Best Text to Pokémon Prompts Best DALL-E Prompts Best Fashion Diffusion Prompts AI Models All AI models Image generation AI Models Video generation AI Models Large Language Models LoRA Textual Inversion ControlNet Hypernetwork Aesthetic Gradient LyCORIS VAE Poses Wildcards Workflows Other Other sites Dreamshot Jobician 2025 © PromptHero @prompthero
1362
FLUX
Learn to create better images with AI! Check out our crash course in prompt engineering & AI art generation! AM1970 posted about 2 hours ago 8 views 0 comments Behold this realistic masterpiece: a slim, elegant woman with fiery red hair tied in a long ponytail, defined eyebrows framing her striking blue eyes, and pale skin dotted with charming freckles across her face and chest. Captured from a side angle, she radiates happiness with a closed-mouth smile, her upturned nose adding to her allure. Dressed in a vibrant, colorful fabric robe pieced together with zips, buttons, seams, and intricate stitching—fully clothed yet teasing partial nudity with subtle cleavage and teardrop-shaped, saggy breasts for a natural, confident vibe. Illuminated by dramatic rim lighting that highlights every textured, detailed inch of her perfect skin, this high-quality image is a celebration of beauty, realism, and artistry! Read more... Comments (0) Add a comment Stunning handsome readhead woman by AM1970 • Follow 0 realistic, masterpiece, high quality BREAK 1woman, 28 y. o. slim fit woman, red hair, long pony tail, defined eyebrows, (fully clothed) , pale skin, (realistic skin:1.3) | (textured skin:1.3) | (detailed skin:1.3) | (perfect skin:1.3) , (upturned nose) , freckles on face and chest, colorful fabric combining to form robe zips, buttons, seams, stitching), partial nudity, teardrop boobs, blue eyes, (cleavage:1.2) , saggy breasts. side angle. happy| closed mouth. ((rim lighting) ) Negative prompt: unrealistic, saturated, high contrast, big nose, painting, drawing, sketch, cartoon, anime, manga, render, CG, 3d, watermark, signature, label, nsfw, nudity, fetish, latex, pvc, suggestive pose, weapon, combat, blood, war scene, aggressive expression, harsh shadows, heavy makeup, color tint, saturation, chromatic aberration, lens flare, low resolution, blur, noise, artifacts, over-retouching, plastic skin, oversharpening, watermark, text, logo Generation parameters 832x1216 Model used CyberRealistic XL v7.0 [Download AI model] More by AM1970 Twitter Instagram Say hi 👋 PromptHero is built by ElevenYellow . Navigation Create AI images Academy - Learn AI PromptHero Community Affiliate program Openjourney AI Challenges Verified profiles Plans Advertise FAQ 🟢 Status Learn AI Content Creation with AI for Brands and Products Midjourney for graphic design & art professionals Crash course in generative AI & prompt engineering for images AI influencers and consistent characters Create custom AI models and LoRas by fine-tuning Stable Diffusion Master your composition: advanced AI image generation with ControlNet Guides Best AI Upscalers 2025 Best AI Video Editors 2025 Best AI 3D Models 2025 Best AI Image Editors 2025 Best AI Notetakers 2025 The Best AI Prompts Best AI Prompts Best Veo Prompts Best Hailuo Prompts Best Wanx Prompts Best Hunyuan Prompts Best Seedance Prompts Best Pixverse Prompts Best Luma Prompts Best Kling Prompts Best Runway Prompts Best Nano Banana Prompts Best SeedEdit Prompts Best Qwen Prompts Best FLUX Prompts Best Recraft Prompts Best Ideogram Prompts Best Seedream 4 Prompts Best Google Imagen Prompts Best ChatGPT Image Prompts Best Stable Diffusion Prompts Best DreamShaper Prompts Best Realistic Vision Prompts Best epiCRealism Prompts Best majicMIX realistic Prompts Best Midjourney Prompts Best Openjourney Prompts Best ChilloutMix Prompts Best Deliberate Prompts Best ReV Animated Prompts Best Perfect World Prompts Best Dreamlike Photoreal Prompts Best Dreamlike Diffusion Prompts Best Anything Prompts Best Protogen Prompts Best AbyssOrangeMix Prompts Best Grapefruit Prompts Best Kenshi Prompts Best Analog Diffusion Prompts Best SynthwavePunk Prompts Best Illuminati Diffusion Prompts Best Arcane Diffusion Prompts Best Mo-Di Diffusion Prompts Best Redshift Diffusion Prompts Best Waifu Diffusion Prompts Best TrinArt Prompts Best Poolsuite Diffusion Prompts Best Funko Diffusion Prompts Best Text to Pokémon Prompts Best DALL-E Prompts Best Fashion Diffusion Prompts AI Models All AI models Image generation AI Models Video generation AI Models Large Language Models LoRA Textual Inversion ControlNet Hypernetwork Aesthetic Gradient LyCORIS VAE Poses Wildcards Workflows Other Other sites Dreamshot Jobician 2025 © PromptHero @prompthero
1226
FLUX
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
7821
Nano Banana
Among the growing list of AI photo editing tools, Deepfake Maker takes a slightly different approach. While it’s best known for facial swaps and animated transformations, it also includes a surprisingly capable AI clothing removal feature. It’s not for everyone, but it’s an interesting choice for users looking for more dramatic or playful results. If you're after a quick way to generate bold clothing edits or outfit swaps, Deepfake Maker might surprise you. It’s not the most refined tool out there, but it’s definitely one of the most entertaining. 4. PxBee PxBee Source: PxBee Available On: Web Browser Price: Free *Annual subscription starts at $00.27/image and you can choose to download 50, 125, or 750 images monthly per year Pros Decent auto-detection for basic clothing removal Integrated with other tools Cons Not suitable for high-end projects Results take longer than expected to generate PxBee is better known for its stock photo collection and image generation features, but it also offers a basic AI clothing removal function built into its photo editing toolkit. While it's not as feature-rich or advanced as some dedicated AI editors, its simplicity and accessibility make it a decent option for casual users who want fast, no-frills results. PxBee may not be the most powerful AI clothing remover out there, but it's a solid pick for quick edits and simple concept visuals. Its intuitive UI and browser-based access make it appealing for users who want fast results without the learning curve. If you're after something basic, affordable, and easy to use, it's a tool worth exploring. 5. AI Face Swap AI Face Swap Source: AI Face Swap Available On: Web Browser Price: $1 per image Pros No manual masking needed Prompt-based customization Presets for inspiration Cons No fine-tuning or editing control With AI Face Swap, you can enter prompts like "red bikini", "formal suit", or even "nude", and the AI will generate a realistic version of your uploaded image that reflects your description. You can also choose from quick preset styles like bikini, romper, or skirt for faster results. This level of control makes the tool especially appealing to users looking for stylized edits, fantasy concepts, or mockup visuals. It also eliminates the need for manual masking or precise editing. 6. Bylo AI BYLO AI Source: Bylo.ai Available On: Web Browser Price: Free *Premium subscription available from $9.99 Pros Prompt-based clothing edits Free trial credits Part of a growing creative suite Cons No manual adjustment tools Limited free usage Relies heavily on prompts Bylo AI is a user-friendly AI platform that offers a range of creative features, including an AI Clothes Remover. The tool is built with simplicity in mind: just upload a photo, enter your prompt, and let the AI generate a clean, edited version. Bylo AI’s Clothes Remover is a great choice if you're looking for a lightweight, affordable tool that lets you experiment with clothing changes using just a text prompt. It won’t replace professional retouching tools, but for concept edits, social media content, or fun transformations, it delivers exactly what most casual users need. FAQs About AI Clothing Remover & Replacer 1. How do I remove clothes in photos? 2. What is the best online tool to remove clothes? 3. Are AI clothes removers safe?
42502
Hero
((abstract illusionism)), 8k full wide focus, (Masterpiece:1.2), ((surreal)),cinematic still, ((photorealistism)), (psychedelic imagery), trending on artstation, sharp focus, studio photo, intricate details, highly detailed, by greg rutkowski In the ethereal realm of the surreal dreamscape, envision a mesmerizing painting that beckons viewers into a kaleidoscope of fantastical wonders, a realm where imagination knows no bounds Negative prompt: nsfw, nude,blurry, low quality, oddly shaped, disfigured,unclear || undefined || mutations || deformities || off center || poor_composition Steps: 100, Sampler: DPM++ 3M SDE, CFG scale: 14.0, Seed: 33, Face restoration: CodeFormer, Size: 1280x832, Model hash: 1895a00bfc, Model: SSD-1B-A1111, Denoising strength: 0.31000000000000005, Mask blur: 4, Noise multiplier: 1.05, Version: v1.6.0-2-g4afaaf8a
0132
Stable Diffusion
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
6516
Nano Banana
<think> The user wants a Z-Image-Turbo prompt based on the provided JSON description. I need to extract information from each field in the JSON and structure it into a prompt following the specified format. Let's break down each section and map it to the prompt structure: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - From "description": stunning model in her early 20s, European features with fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on arms and cheeks, natural micro-imperfections, light natural freckles across nose and upper chest, lightly tanned sun-kissed glow on collarbones and shoulders, athletic yet feminine body with realistic natural proportions, medium perky breasts with soft realistic shape and subtle natural cleavage visible through slightly sheer sections of her form-fitting white and black tactical outfit, low-cut neckline and tight fit accentuating toned midriff, slim waist, defined hips and long legs. - From "face": Long tousled blonde hair with natural volume, sun-kissed highlights and individual strands moving gently as she walks, extremely detailed large beautiful brilliant blue-silver-golden-grey eyes with intricate iris textures, realistic reflections and natural depth, eyes fluttering while staring directly into the camera with intense captivating gaze, full glossy lips transitioning into a warm radiant genuine smile over the scene showing perfect white teeth, high cheekbones, soft natural blush on cheeks. - From "outfit": Form-fitting white and black tactical outfit with futuristic design: high-tech fabric with subtle matte-gloss contrast, deep V-neckline and strategic sheer mesh panels revealing skin tone and natural curves underneath for subtle NSFW sensuality, tactical straps and utility details on shoulders and hips, black accents on sleeves and waist, no visible underwear lines for clean realistic look. 2. **Action or scene context** - From "pose": Dynamic confident walk down a futuristic corridor, hips swinging gracefully, one leg forward in mid-stride, arms relaxed yet purposeful, head slightly tilted toward camera with engaging direct eye contact and growing smile. 3. **Environment / setting + time of day + weather** - From "environment": Sleek futuristic corridor with metallic walls reflecting soft light, continuous yellow glowing floor line creating strong perspective and depth, vertical warm amber light panels on walls casting cinematic illumination, subtle sci-fi ambient fog in distance, high-tech yet elegant atmosphere. 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - From "lighting": Warm diffused lighting from vertical panels on both sides, soft golden-amber glow blending with cool metallic reflections, gentle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing three-dimensional form and skin texture, high dynamic range with vivid yet natural colors. 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - From "camera": Cinematic full-body shot with medium telephoto lens (85–135mm equivalent), f/2.0 aperture, razor-sharp focus from face to feet, shallow depth of field with creamy smooth bokeh on corridor background, slight low angle to emphasize power and grace, perfect viewpoint. 6. **Mood / atmosphere / emotion** - From "style" and "environment": elegant futuristic atmosphere with graceful and seductive vibe. 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - From "style": Ultra photorealistic cinematic photography. - From "quality": trending professional sci-fi fashion portrait. 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - From "quality": 8K resolution, extreme sharpness, rich cinematic colors with warm amber tones and cool metallic contrast, highly detailed skin realism, hair movement, outfit texture, corridor reflections and light play. - From "style" and general understanding: masterpiece quality. Now, assemble the prompt in the specified format. </think> A stunning young woman with European features and fair smooth skin showcasing hyper-detailed realistic textures, including visible fine pores, subtle peach fuzz, natural micro-imperfections, light freckles, sun-kissed glow on collarbones, athletic yet feminine body with realistic proportions, medium perky breasts with subtle natural cleavage visible through sheer sections of a form-fitting white and black tactical outfit, low-cut neckline accentuating toned midriff, slim waist, defined hips, long legs, long tousled blonde hair with sun-kissed highlights and individual strands gently moving, extremely detailed large brilliant blue-silver-golden-grey eyes with intricate iris textures and realistic reflections, full glossy lips transitioning into a warm radiant smile showing perfect white teeth, high cheekbones, soft natural blush on cheeks, dynamic confident walk down a futuristic corridor with hips swinging gracefully, one leg forward in mid-stride, arms relaxed yet purposeful, head tilted toward camera with engaging direct eye contact and growing smile, sleek futuristic corridor with metallic walls reflecting soft light, continuous yellow glowing floor line creating perspective and depth, vertical warm amber light panels on walls casting cinematic illumination, subtle sci-fi ambient fog in distance, high-tech yet elegant atmosphere, warm diffused lighting from vertical panels, soft golden-amber glow blending with cool metallic reflections, gentle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing form and skin texture, high dynamic range with vivid natural colors, cinematic full-body shot with medium telephoto lens (85–135mm equivalent), f/2.0 aperture, razor-sharp focus from face to feet, shallow depth of field with creamy smooth bokeh on corridor background, slight low angle emphasizing power and grace, perfect viewpoint, ultra photorealistic cinematic photography, maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture, realistic hair strands with natural movement, authentic fabric weave and subtle sheer effect, elegant futuristic atmosphere with graceful and seductive vibe, 8K resolution, extreme sharpness, rich cinematic colors with warm amber tones and cool metallic contrast, highly detailed skin realism, hair movement, outfit texture, corridor reflections and light play, trending professional sci-fi fashion portrait, masterpiece quality, ultra detailed, absurdres, intricate, sharp focus, depth of field, best quality, 8k uhd, hdr, photorealistic, realistic textures, dramatic rim lighting, volumetric fog, golden hour lighting, The subject is sharply depicted with outstanding detail., flux_pro
201525
Z Image
Among the growing list of AI photo editing tools, Deepfake Maker takes a slightly different approach. While it’s best known for facial swaps and animated transformations, it also includes a surprisingly capable AI clothing removal feature. It’s not for everyone, but it’s an interesting choice for users looking for more dramatic or playful results. If you're after a quick way to generate bold clothing edits or outfit swaps, Deepfake Maker might surprise you. It’s not the most refined tool out there, but it’s definitely one of the most entertaining. 4. PxBee PxBee Source: PxBee Available On: Web Browser Price: Free *Annual subscription starts at $00.27/image and you can choose to download 50, 125, or 750 images monthly per year Pros Decent auto-detection for basic clothing removal Integrated with other tools Cons Not suitable for high-end projects Results take longer than expected to generate PxBee is better known for its stock photo collection and image generation features, but it also offers a basic AI clothing removal function built into its photo editing toolkit. While it's not as feature-rich or advanced as some dedicated AI editors, its simplicity and accessibility make it a decent option for casual users who want fast, no-frills results. PxBee may not be the most powerful AI clothing remover out there, but it's a solid pick for quick edits and simple concept visuals. Its intuitive UI and browser-based access make it appealing for users who want fast results without the learning curve. If you're after something basic, affordable, and easy to use, it's a tool worth exploring. 5. AI Face Swap AI Face Swap Source: AI Face Swap Available On: Web Browser Price: $1 per image Pros No manual masking needed Prompt-based customization Presets for inspiration Cons No fine-tuning or editing control With AI Face Swap, you can enter prompts like "red bikini", "formal suit", or even "nude", and the AI will generate a realistic version of your uploaded image that reflects your description. You can also choose from quick preset styles like bikini, romper, or skirt for faster results. This level of control makes the tool especially appealing to users looking for stylized edits, fantasy concepts, or mockup visuals. It also eliminates the need for manual masking or precise editing. 6. Bylo AI BYLO AI Source: Bylo.ai Available On: Web Browser Price: Free *Premium subscription available from $9.99 Pros Prompt-based clothing edits Free trial credits Part of a growing creative suite Cons No manual adjustment tools Limited free usage Relies heavily on prompts Bylo AI is a user-friendly AI platform that offers a range of creative features, including an AI Clothes Remover. The tool is built with simplicity in mind: just upload a photo, enter your prompt, and let the AI generate a clean, edited version. Bylo AI’s Clothes Remover is a great choice if you're looking for a lightweight, affordable tool that lets you experiment with clothing changes using just a text prompt. It won’t replace professional retouching tools, but for concept edits, social media content, or fun transformations, it delivers exactly what most casual users need. FAQs About AI Clothing Remover & Replacer 1. How do I remove clothes in photos? 2. What is the best online tool to remove clothes? 3. Are AI clothes removers safe?
31317
Hero
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
5393
Nano Banana
{ "prompt_type": "descriptive_replication", "reference_adherence": "STRICT_VISUAL_FIDELITY", "aspect_ratio": "3:4", "identity_lock": { "priority": "NONE", "instruction": "New female model. Completely different facial structure and overall appearance from any previous version. No resemblance to prior identity." }, "scene_constraints": { "people_count": "exactly one person in the frame", "single_subject_only": true, "no_other_people_visible": true, "no_extra_reflections": true, "no_background_figures": true, "full_body_visible": true, "head_to_toe_in_frame": true }, "subject": { "demographics": "young woman (age 20+), blonde", "presence": "only one visible person in the entire scene", "hair": { "color": "warm blonde (no dark roots)", "style": "side part, long straight hair", "texture": "smooth, natural fall, lightly voluminous" }, "face": { "structure": "distinct oval face shape with higher cheekbones and a slightly sharper jawline", "expression": "calm alluring expression with a subtle closed-lip smirk", "eyes": "almond-shaped eyes with a slightly lifted outer corner", "gaze": "looking at the phone screen through the mirror with composed confidence", "nose": "straight narrow nose with refined bridge", "lips": "naturally fuller lips with soft definition", "makeup": "minimal natural makeup, soft matte skin finish, muted rose lips, softly blended eyeshadow", "visibility": "partially covered by phone but clearly a different facial identity than any previous version" } }, "body": { "pose": "standing in front of mirror with one arm raised above head, legs naturally extended", "posture": "relaxed confident stance with slight hip shift", "anatomy": { "curves": "natural soft curves with realistic proportions", "chest": "bikini-supported, natural shape", "details": "visible collarbones, subtle shoulder definition, natural leg proportions" }, "feet": { "visibility": "fully visible at bottom of frame", "position": "flat on floor, natural stance" }, "skin_texture": "natural skin texture, no airbrushing" } }, "attire_and_accessories": { "clothing": { "top": "black triangle string bikini top with thin straps", "bottom_detail": "matching bikini bottom strings visible at hips" }, "footwear": "barefoot", "jewelry": { "necklaces": [], "wrists": [], "fingers": "no rings" }, "props": { "phone": "white iPhone (iPhone Pro style), held in right hand" } }, "environment": { "setting": "modern luxury bathroom interior, daytime", "foreground_element": { "mirror": "large rectangular mirror with GREY MARBLE FRAME", "material": "polished stone marble with subtle veining" }, "background": { "elements": [ "large marble wall tiles", "floating vanity", "minimalist sink and chrome fixtures", "folded white towels", "clean tiled floor fully visible" ], "style": "minimalist and clean", "clutter": "none" } }, "lighting_and_atmosphere": { "source": "natural daylight from camera-left", "quality": "soft directional light", "effects": [ "subtle highlights on shoulder and collarbone", "gentle shadow falloff along legs and feet" ], "contrast": "moderate, natural" }, "camera_and_technical": { "perspective": "mirror selfie", "camera_position": "eye level, slightly angled downward", "framing": "vertical 3:4, full head-to-toe body visible without cropping feet", "focus": "sharp on subject, background slightly softer", "visual_fidelity": "real smartphone camera look, mild natural noise" }, "realism_constraints": { "allowed": [ "natural skin texture", "minor asymmetry", "realistic lighting" ], "forbidden": [ "multiple people", "extra person", "background people", "beauty filters", "airbrushed skin", "anime style", "cartoon style", "cropped feet", "cut-off legs" ] } }
1356
Nano Banana
{ "generation_request": { "meta_data": { "tool": "NanoBanana Pro", "task_type": "photobooth_contact_sheet_fashion_grid", "version": "v1.1_REFERENCE_IDENTITY_GRID_EN", "priority": "highest", "language": "en" }, "input": { "mode": "image_to_image", "reference_image_usage": "required", "notes": "Use the uploaded reference image to preserve the same person’s face, body features, hair, makeup and clothing. Maintain identity consistency in all frames." }, "output_settings": { "aspect_ratio": "4:5", "orientation": "portrait", "resolution_target": "ultra_high_res", "num_images": 1, "layout": { "type": "grid", "rows": 4, "cols": 4, "gutter": "thin_black_borders", "outer_border": "black_frame" }, "sharpness": "slightly_soft_flash", "film_grain": "medium_analog", "color_grade": "warm_2000s_flash_tone", "dynamic_range": "natural_not_hdr" }, "creative_prompt": { "scene_summary": "Early 2000s disposable camera photobooth contact sheet featuring the same person from the reference image in all 16 frames. Playful, candid, flash-lit aesthetic.", "environment": { "background": "plain white wall indoors", "lighting": "direct on-camera flash, slight hard shadow behind subject, subtle hotspot on skin, authentic point-and-shoot look" }, "subject": { "identity_consistency": "exact same person as reference image in all frames", "hair": "match reference image exactly", "makeup": "match reference image exactly", "wardrobe": "preserve the exact clothing from the reference image with accurate fabric texture and fit", "accessories": "keep accessories consistent with reference image" }, "poses_and_expressions": [ "hands on hips confident pose", "soft pout toward camera", "gentle smile", "hand on chin thoughtful look", "playful wink", "hands in hair", "slight shoulder turn", "looking off to the side", "serious model face", "soft laugh", "blowing a kiss", "tilted head cute expression", "crossed arms relaxed pose", "side glance over shoulder", "chin resting on hand", "natural relaxed expression" ], "style": { "aesthetic": "2000s photobooth disposable camera", "texture": "visible film grain, dust particles, tiny scratches, mild vignetting", "color": "slightly warm faded tones, subtle flash flattening", "lens": "35mm point-and-shoot softness" } }, "negative_prompt": [ "text", "logos", "watermark", "different outfit between frames", "different hairstyle", "different person", "face distortion", "identity drift", "overly polished editorial look", "HDR" ] } }
391
Among the growing list of AI photo editing tools, Deepfake Maker takes a slightly different approach. While it’s best known for facial swaps and animated transformations, it also includes a surprisingly capable AI clothing removal feature. It’s not for everyone, but it’s an interesting choice for users looking for more dramatic or playful results. If you're after a quick way to generate bold clothing edits or outfit swaps, Deepfake Maker might surprise you. It’s not the most refined tool out there, but it’s definitely one of the most entertaining. 4. PxBee PxBee Source: PxBee Available On: Web Browser Price: Free *Annual subscription starts at $00.27/image and you can choose to download 50, 125, or 750 images monthly per year Pros Decent auto-detection for basic clothing removal Integrated with other tools Cons Not suitable for high-end projects Results take longer than expected to generate PxBee is better known for its stock photo collection and image generation features, but it also offers a basic AI clothing removal function built into its photo editing toolkit. While it's not as feature-rich or advanced as some dedicated AI editors, its simplicity and accessibility make it a decent option for casual users who want fast, no-frills results. PxBee may not be the most powerful AI clothing remover out there, but it's a solid pick for quick edits and simple concept visuals. Its intuitive UI and browser-based access make it appealing for users who want fast results without the learning curve. If you're after something basic, affordable, and easy to use, it's a tool worth exploring. 5. AI Face Swap AI Face Swap Source: AI Face Swap Available On: Web Browser Price: $1 per image Pros No manual masking needed Prompt-based customization Presets for inspiration Cons No fine-tuning or editing control With AI Face Swap, you can enter prompts like "red bikini", "formal suit", or even "nude", and the AI will generate a realistic version of your uploaded image that reflects your description. You can also choose from quick preset styles like bikini, romper, or skirt for faster results. This level of control makes the tool especially appealing to users looking for stylized edits, fantasy concepts, or mockup visuals. It also eliminates the need for manual masking or precise editing. 6. Bylo AI BYLO AI Source: Bylo.ai Available On: Web Browser Price: Free *Premium subscription available from $9.99 Pros Prompt-based clothing edits Free trial credits Part of a growing creative suite Cons No manual adjustment tools Limited free usage Relies heavily on prompts Bylo AI is a user-friendly AI platform that offers a range of creative features, including an AI Clothes Remover. The tool is built with simplicity in mind: just upload a photo, enter your prompt, and let the AI generate a clean, edited version. Bylo AI’s Clothes Remover is a great choice if you're looking for a lightweight, affordable tool that lets you experiment with clothing changes using just a text prompt. It won’t replace professional retouching tools, but for concept edits, social media content, or fun transformations, it delivers exactly what most casual users need. FAQs About AI Clothing Remover & Replacer 1. How do I remove clothes in photos? 2. What is the best online tool to remove clothes? 3. Are AI clothes removers safe?
2677
Hero
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
3246
Nano Banana
--- name: pastilles-llm-multi-agent description: Variante Claude Code du générateur de pastilles LLM, qui lance réellement plusieurs sous-agents en parallèle. Génère une pastille de communication interne sur les LLM (un texte court en français plus un prompt unique de génération d'images à coller dans le chat Gemini), via cinq sous-agents indépendants, une fusion, puis une revue critique par trois sous-agents et une correction. Utilise ce skill dès qu'on te demande de rédiger, produire ou générer une pastille, une fiche ou un contenu court d'acculturation sur les LLM, l'IA générative, le prompting, les agents, le RAG, la confidentialité IA ou tout sujet de la liste des 45 pastilles ci-dessous, que le mot "pastille" soit employé ou non. Utilise-le aussi dès qu'on te fournit un titre issu de cette liste. --- # Générateur de pastilles LLM, version multi-agents (Claude Code) ## Ce que fait ce skill Produit une pastille pédagogique complète à partir d'un titre. À la différence de la version chat, il lance de vrais sous-agents parallèles: une passe de recherche, puis cinq sous-agents indépendants qui rédigent chacun un brouillon sous un angle différent, puis une fusion par l'orchestrateur qui retient les meilleures formulations, et enfin une revue critique par trois sous-agents suivie d'une correction. Livrable: le texte de la pastille et un prompt unique de génération d'images à coller dans le chat Gemini. ## Environnement requis Ce skill suppose Claude Code, avec les sous-agents (outil Task) et la recherche web disponibles. Si les sous-agents ne sont pas disponibles dans ton environnement, n'improvise pas: utilise la variante chat, qui fait le même travail en self-ensemble séquentiel. ## Entrée Un titre de pastille, idéalement issu de la liste ci-dessous. Si le titre est ambigu ou hors liste, demande une clarification avant de générer. ## Liste des 45 pastilles (pour la continuité) 1. Au fait, c'est quoi un LLM ? 2. Modèles de fondation : ces géants pré-entraînés dont tout découle 3. Dans les coulisses : comment une IA apprend (sans vraiment comprendre) 4. Le contexte : la "mémoire vive" de l'IA (ce qu'elle voit à un instant T) 5. Pourquoi l'IA oublie : la vérité sur la mémoire d'une session à l'autre 6. Température et créativité : régler le curseur de la "folie" de l'IA 7. Les tokens : la monnaie d'échange (et la manière de penser) des LLM 8. Le non-déterminisme : pourquoi l'IA fait des stats et ne donne jamais deux fois la même réponse 9. Multimodalité : quand l'IA commence à voir, entendre et coder 10. Knowledge cutoff : pourquoi votre IA vit (parfois) dans le passé 11. Les biais de l'IA : le miroir pas toujours glorieux de nos propres données 12. Zéro pointé : pourquoi l'IA est structurellement nulle en calcul mental 13. Les hallucinations : quand l'IA invente la vie avec un aplomb incroyable 14. L'art du Fact-Checking : comment ne pas gober les hallucinations de l'IA 15. L'empreinte carbone de l'IA : combien d'eau et de CO2 pour un prompt ? 16. Anatomie d'un bon prompt : la recette de base 17. Les pièges du prompt : flou artistique et overdose d'instructions 18. L'art de la discussion : pourquoi il faut itérer plutôt que tout demander d'un coup 19. Donner des ordres : comment imposer un format de sortie strict (JSON, tableaux, etc.) 20. Diviser pour régner : découper une tâche complexe en étapes claires 21. Le Few-Shot Prompting : un bon exemple vaut mille consignes 22. Inception : donner un rôle et un contexte à votre IA 23. Zéro syndrome de la page blanche : rédiger et reformuler ses mails ou CR 24. La diagonale du fou : résumer et synthétiser un pavé de 50 pages en 2 minutes 25. Le sparring-partner : brainstormer et structurer ses idées sans filtre 26. Caméléon : traduire et adapter son ton selon le destinataire 27. Data-analyst de poche : faire parler un tableur ou des données brutes 28. Mise en situation : préparer une réunion ou un brief client avec l'IA 29. Avis aux dévs : l'IA comme copilote de code et de review 30. SLM : Pourquoi sortir un tank pour écraser une mouche ? 31. Le "harnais" : donner des bras, des yeux et un cadre au LLM 32. C'est quoi un "Agent" IA, concrètement ? 33. Human-in-the-loop : pourquoi il faut toujours garder un humain aux manettes 34. Quand l'agent s'emballe : autonomie et erreurs en cascade 35. Automatisation : confier un workflow complet à un agent 36. Tool Use : quand l'IA apprend à utiliser une calculatrice ou une API 37. Le RAG : faire réviser vos propres documents à l'IA avant qu'elle ne réponde 38. MCP et connecteurs : brancher l'IA directement sur vos outils de dev et d'orga 39. Responsabilité : vous êtes le seul signataire de ce que produit l'IA 40. Confidentialité : où vont vraiment les données que vous tapez ? 41. Alerte Rouge : les données clients à ne JAMAIS copier-coller dans une IA public 42. Shadow AI : pourquoi le ChatGPT gratuit perso au bureau est une fausse bonne idée 43. RGPD et IA : ce qu'il faut savoir pour rester dans les clous 44. Prompt Injection : quand on pirate une IA avec de simples phrases 45. La Chaîne de Pensée : pourquoi l'IA a besoin de "réfléchir à voix haute" Sers-toi de cette liste pour situer la pastille et délimiter son périmètre. Avant de rédiger: - Repère les 1 à 3 pastilles voisines qui recouvrent le sujet, et dresse une courte liste "déjà traité ailleurs, à ne pas ré-expliquer". - Si une pastille voisine a déjà été produite, demande son texte à l'utilisateur (ou appuie-toi dessus s'il est fourni) pour caler précisément la limite. - Les concepts fondateurs déjà couverts ne sont rappelés qu'en une phrase de mise en contexte, jamais ré-expliqués. Le coeur de la pastille est consacré à ce que son titre ajoute de neuf. - Ne renvoie pas vers les autres pastilles dans le texte final, sauf si le titre l'impose. ## Règles du texte - Longueur adaptée à la profondeur du sujet. Un sujet léger tient en 3 paragraphes courts; un sujet plus riche peut aller jusqu'à 4 paragraphes plus étoffés. Ne gonfle pas artificiellement un sujet simple et ne compresse pas à l'excès un sujet dense: juge la profondeur par la richesse réelle du concept. Prose uniquement, pas de listes ni de puces. - Ton décontracté, précis et léger: accessible, vivant et sans lourdeur, mais techniquement juste. Public visé: une main-d'oeuvre diverse en société de conseil, du profil non technique au développeur. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Reste léger, évite le name-dropping. N'accumule pas les noms de modèles, d'outils, d'entreprises, de chercheurs ou de techniques pour faire savant: privilégie l'explication claire du mécanisme. Ne cite un nom précis que s'il éclaire vraiment le propos, une ou deux fois au maximum. La recherche sert l'exactitude et alimente la section Sources, elle n'a pas à truffer le texte de références. - Contenu autonome: la pastille se comprend seule, sans avoir lu les autres, sauf mention contraire dans le titre. - Rédige en français. - Ne mentionne jamais ANEO ni aucun nom d'entreprise. - Le texte explicatif ne figure jamais dans l'image. - N'utilise pas de tiret cadratin. Privilégie des caractères standard (virgules, deux-points, parenthèses). ## Processus multi-agents ### Étape 1, recherche (orchestrateur, une seule fois) Toi, l'orchestrateur, fais d'abord une passe de recherche web ciblée sur le sujet de la pastille. Une seule passe, en amont. Les sous-agents ne peuvent pas lancer de sous-agents et n'héritent pas de ton contexte, donc c'est à toi de faire la recherche puis de leur transmettre les résultats. - Objectif: exactitude, chiffres, exemples concrets et faits à jour. - Ancre tes recherches dans le présent: repère la date du jour fournie dans ton contexte (champ currentDate) et privilégie les informations et chiffres les plus récents, car ce domaine évolue vite. Précise cette date au besoin dans tes requêtes. - Priorise les sources originales ou officielles (blogs d'éditeurs, articles de recherche, sites gouvernementaux, documentation) plutôt que les agrégateurs. - Adapte la profondeur: quelques recherches pour un concept stable (par exemple ce qu'est un token), davantage pour un sujet à chiffres ou mouvant (empreinte carbone et coûts, modèles spécifiques, SLM, MCP et connecteurs, RGPD). - Produis un brief de recherche compact: faits clés, chiffres utiles, et 2 à 4 sources. Tu l'injecteras tel quel dans chaque sous-agent, et tu listeras les sources dans le livrable final. ### Étape 2, affinement de la thématique (orchestrateur, une seule fois) Si plusieurs axes apparaissent pertinents après avoir enlever les axes adressés par d'autres pastilles prévues, dialogue avec l'utilisateur pour choisir l'axe de la pastille. Si l'axe choisi nécessite de nouvelles sources, refait en une recherche web selon les consignes précédentes. ### Étape 3, fan-out (cinq sous-agents en parallèle) Lance cinq sous-agents via l'outil Task, dans le même tour, pour qu'ils s'exécutent en parallèle. Sois explicite sur le parallélisme: par défaut Claude Code reste séquentiel, il ne parallélise que si tu le demandes clairement. Un sous-agent par angle: 1. Analogie: expliquer le concept via une métaphore concrète du quotidien. 2. Cas d'usage: partir d'une situation de travail réelle et vécue. 3. Idée reçue: démarrer d'un malentendu courant, puis rectifier. 4. Mécanique: aller au coeur du fonctionnement, précis mais accessible. 5. Enjeu: pourquoi ça compte, ce que ça change concrètement pour l'utilisateur. Comme le contexte n'est pas hérité, chaque prompt de sous-agent doit être autonome et contenir tout le nécessaire. Utilise ce gabarit, en remplaçant les crochets: ``` Tu rédiges un brouillon d'une pastille pédagogique interne sur les LLM. Tu travailles seul, sans accès au reste de la conversation. Titre de la pastille: [TITRE] Angle imposé: [ANGLE, par exemple "Analogie: expliquer via une métaphore concrète du quotidien"] À NE PAS ré-expliquer (déjà couvert par d'autres pastilles), à mentionner en une phrase au maximum: [LISTE des points déjà traités ailleurs, à ne pas retraiter] Brief de recherche (appuie-toi dessus, ne relance pas de recherche): [BRIEF: faits clés, chiffres, sources] Règles: - Longueur selon la profondeur du sujet: 3 paragraphes courts si le sujet est léger, jusqu'à 4 paragraphes étoffés s'il est dense. Prose uniquement, pas de listes. - Mets en gras et/ou en italique les termes importants, de l'ordre de 3 à 4 emphases par paragraphe (repère indicatif, ne force pas ce compte). - Ton décontracté, précis et léger, accessible mais techniquement juste. Évite le name-dropping: pas d'accumulation de noms de modèles, outils, entreprises ou chercheurs, privilégie l'explication du mécanisme. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Français. Ne mentionne aucune entreprise. Pas de tiret cadratin. - Contenu autonome, compréhensible seul. Réponds exactement dans ce format: TEXTE: [les 3 à 4 paragraphes] ILLUSTRATION_TITRE: [une phrase décrivant le concept central à illustrer] SCHEMA: [oui ou non. Si oui, décris en 2 à 3 lignes le diagramme et ses libellés en français] ``` ### Étape 4, fan-in et fusion (orchestrateur) Attends les cinq retours, puis fusionne en une seule pastille finale: - Retiens la structure la plus claire, la meilleure analogie, l'exemple le plus parlant, la correction la plus nette, la formulation la plus précise et le "pourquoi ça compte" le plus fort. - Réécris en une seule voix cohérente. Pas d'effet patchwork. - Respecte la longueur adaptée à la profondeur (voir Règles du texte) et le ton décontracté, précis et léger, sans name-dropping. - Décide s'il faut un schéma: inclus-en un si au moins deux sous-agents ont répondu "oui", ou si le concept est intrinsèquement un processus, un flux ou une comparaison (par exemple RAG, agents, étapes de prompting, comparaison de modèles). Le cas échéant, fusionne les meilleures idées de schéma. - Construis ensuite le prompt image unique (section ci-dessous). ### Règles d'écriture pour la pastille finale - Longueur selon la profondeur du sujet: 3 paragraphes courts si le sujet est léger, jusqu'à 4 paragraphes étoffés s'il est dense. Prose uniquement, pas de listes. - Mets en gras et/ou en italique les termes importants, de l'ordre de 3 à 4 emphases par paragraphe (repère indicatif, ne force pas ce compte). - Ton décontracté, précis et léger, accessible mais techniquement juste. Évite le name-dropping: pas d'accumulation de noms de modèles, outils, entreprises ou chercheurs, privilégie l'explication du mécanisme. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Français. Ne mentionne aucune entreprise. Pas de tiret cadratin. - Contenu autonome, compréhensible seul. ### Étape 5, revue critique et correction (trois sous-agents en parallèle, puis orchestrateur) Quand la lancer: systématiquement à la première génération, une fois le texte fusionné et le prompt image construits. Ensuite, si l'utilisateur demande des ajustements, ne relance pas la revue d'office: propose-la, et ne la lance qu'avec son accord (elle coûte trois sous-agents de plus). Principe: les relecteurs produisent des CONSTATS, jamais une réécriture. Chacun renvoie une liste de problèmes localisés, assortis d'une gravité et d'un correctif précis. C'est toi, l'orchestrateur, qui appliques et réécris, pour garder une voix unique et éviter l'effet patchwork. Ce que la revue peut juger: les relecteurs ne voient pas les images, générées plus tard dans Gemini. La revue du visuel porte donc uniquement sur le PROMPT image (clarté, cohérence avec le texte, respect des consignes), jamais sur un rendu. Le contrôle visuel réel reste à l'utilisateur. Lance trois sous-agents en parallèle (parallélisme explicite, sinon l'exécution sera séquentielle), un par grille. Le contexte n'étant pas hérité, chaque prompt doit être autonome et contenir tout le nécessaire: le titre, le texte fusionné, le bloc prompt image, le brief de recherche, la liste "déjà traité ailleurs" (et les textes des pastilles voisines s'ils sont disponibles), et la liste des 45 titres pour repérer les chevauchements. Les trois grilles: 1. Fond, exactitude et périmètre: exactitude vs le brief (aucun chiffre, date ou fait inventé ni sur-affirmé, rien qui le contredise), cohérence entre le titre et ce que le texte délivre, chevauchements avec les pastilles voisines (le texte ré-explique-t-il ce qui est traité ailleurs ? redites à signaler), autonomie du contenu, clarté du message à retenir, repérage de ce qui vieillira mal (à nuancer). 2. Forme, ton et pédagogie: ton décontracté-précis-léger et techniquement juste, absence de name-dropping, rythme et fluidité à la lecture à voix haute, longueur adaptée à la profondeur (3 à 4 paragraphes, prose, pas de listes) et chasse au verbiage, accessibilité pour un profil non technique (jargon expliqué, analogies claires), présence d'emphases utiles (gras/italique), de l'ordre de 3 à 4 par paragraphe à titre indicatif (signale un texte trop peu emphasé autant qu'un excès, mais ne fais jamais retirer une emphase pertinente), force de l'accroche. 3. Conformité et visuel: contraintes dures (aucun tiret cadratin ni caractère non standard, aucun nom d'entreprise ni ANEO, français correct, prose sans listes ni puces); puis le prompt image: titre exact au caractère près, charte présente une seule fois, texte de la pastille bien marqué "à ne pas afficher", illustration-titre iconique et non schéma de processus, schéma en 2e image séparée si retenu, libellés de schéma courts (groupes nominaux) et en français, cohérence entre le visuel décrit et le coeur du texte. Gabarit de relecteur (remplace les crochets): ``` Tu relis un brouillon quasi final de pastille pédagogique interne sur les LLM. Tu travailles seul, sans accès au reste de la conversation. Tu ne réécris pas: tu rends des constats et des correctifs précis. Titre de la pastille: [TITRE] Texte fusionné à relire: [TEXTE COMPLET] Prompt image associé (à relire aussi): [BLOC PROMPT IMAGE] Brief de recherche de référence (fait foi pour l'exactitude): [BRIEF] Déjà traité par d'autres pastilles, à ne pas ré-expliquer (et textes voisins si fournis): [LISTE + TEXTES VOISINS] Liste des 45 titres (pour repérer les chevauchements): [LISTE DES 45 TITRES] Ta grille de relecture (tiens-t'en strictement à elle): [RUBRIQUE 1, 2 OU 3] Rappel: tu ne peux pas voir d'image; juge seulement le PROMPT image, pas un rendu. Réponds exactement dans ce format: CONSTATS: - [bloquant|recommandé|mineur] [problème localisé] -> [correctif précis] - ... (n'invente pas de problème: si un point est bon, ne le liste pas) VERDICT: [une phrase: publiable tel quel / corrections mineures / corrections nécessaires] ``` Fan-in et correction (orchestrateur): rassemble les constats, dédoublonne, arbitre les conseils contradictoires. Garde-fou anti-gonflement: entre "ajouter" et "raccourcir", la concision l'emporte, sauf erreur de fond avérée. Applique les constats bloquants et recommandés, écarte ou mentionne les mineurs. Une seule passe, pas de boucle. Réécris la version corrigée en une seule voix, puis prépare le court résumé "Ce que la revue a corrigé" (voir Format de sortie). ### Points de vigilance Claude Code - Parallélisme explicite: demande clairement cinq sous-agents en parallèle, sinon l'exécution sera séquentielle. - Contexte non hérité: mets tout dans le prompt du sous-agent, en particulier le brief de recherche. - Pas d'imbrication: un sous-agent ne peut pas en lancer un autre, garde recherche et fusion chez l'orchestrateur. - Coût et limites: cinq sous-agents en parallèle multiplient l'usage de jetons et peuvent déclencher des limites de débit. En cas de limite, replie-toi sur une exécution séquentielle. - Revue: la phase de revue ajoute trois sous-agents. Elle tourne d'office à la première génération; sur les demandes d'ajustement ultérieures, propose-la et ne la lance qu'avec l'accord de l'utilisateur. En cas de limite de débit, exécute les relecteurs en séquentiel ou replie-toi sur un relecteur global unique. - Réemploi optionnel: tu peux définir les cinq angles comme sous-agents personnalisés dans .claude/agents/ pour les réutiliser, mais ce n'est pas obligatoire, les sous-agents polyvalents suffisent. ## Prompt de génération d'images (un seul bloc, à coller dans le chat Gemini) Produis un seul bloc de prompt, rédigé en français, à coller tel quel dans le chat de Gemini. Gemini se charge d'appeler Nano Banana une ou deux fois, ce qui réduit les manipulations pour l'utilisateur. Le bloc demande: - Toujours une illustration-titre. - Un schéma seulement s'il a été retenu à l'étape de fusion, généré comme une seconde image séparée (fichier distinct), jamais dans la même image que le titre. Contenu du bloc: - Décris la charte graphique une seule fois (bloc ci-dessous) et précise que toutes les images la partagent, pour une cohérence visuelle. - Précise: tout texte affiché dans les images est en français. Cette consigne est importante, elle évite qu'un libellé se retrouve en anglais. - Inclus le texte complet de la pastille dans le bloc, en contexte de génération, clairement marqué comme à ne pas afficher. Il aide le modèle à comprendre le sujet et à choisir une illustration juste. Les seuls textes rendus visibles sont le titre exact et, pour un schéma, ses libellés. Le texte de la pastille ne doit jamais apparaître dans les images. - Illustration-titre: composition épurée et moderne, focus graphique central iconique représentant le concept. Ce n'est pas un schéma de processus. Le seul texte affiché est le titre exact, en en-tête: n'ajoute aucun sous-titre, accroche ou texte secondaire, le titre seul. Technique en deux temps pour le titre: dans le prompt généré, écris le titre exact entre guillemets droits (forme attendue: Le titre exact: "...") et demande un rendu fidèle, sans faute, en police sans serif corporate. Les guillemets ne sont qu'un délimiteur côté prompt: ne les commente pas et n'ajoute aucune consigne à leur sujet dans le prompt généré (pas de mention du type "les guillemets ne doivent pas apparaître"). - Schéma s'il est retenu: une seconde image générée séparément, dans son propre fichier, jamais fusionnée avec l'illustration-titre ni juxtaposée dans la même image. Diagramme d'entreprise propre et net (processus, flowchart ou comparaison), cohérent avec l'illustration-titre, mêmes charte et style, libellés fournis explicitement en français, très peu de fioritures. - Libellés de schéma: des groupes nominaux courts et lisibles seuls (par exemple "Exemples de bonnes réponses"), pas des phrases verbales qui sonnent comme des ordres ("Montrer de bonnes réponses"). Quelques mots par libellé. Attention aux titres longs: Nano Banana rend fiablement les libellés courts mais peut faire une faute sur un titre long; la technique en deux temps limite le risque. Si le rendu d'un titre long reste incertain, indique à l'utilisateur qu'il peut demander à Gemini d'utiliser Nano Banana Pro, plus fiable pour le texte long. Gabarit, illustration-titre seule (exemple avec le titre 1): ``` Génère une image en respectant cette charte graphique: [bloc Charte graphique]. Contexte pour comprendre le sujet, à NE PAS afficher dans l'image: [ici le texte complet de la pastille]. Illustration-titre pour une pastille pédagogique sur les LLM. Composition épurée et moderne, focus graphique central iconique: un nuage de lettres et de mots qui se condense en une bulle de dialogue, évoquant un modèle de langage qui transforme du texte en réponse. Ce n'est pas un schéma de processus. Seul texte à afficher dans l'image: le titre, en en-tête, sans faute d'orthographe, en police sans serif corporate, sans sous-titre ni texte secondaire. Le titre exact: "TITRE EXACT". Tout le texte affiché est en français. Format 16:9. ``` Gabarit, illustration-titre plus schéma (deux images): ``` Prépares-toi à générer deux images séparées, dans deux fichiers distincts. Les deux partagent cette charte graphique: [bloc Charte graphique]. Tout texte affiché dans les images est en français. Contexte pour comprendre le sujet, à NE PAS afficher dans les images: [ici le texte complet de la pastille]. Image 1, illustration-titre: composition épurée et moderne, focus graphique central iconique qui illustre le sujet. Ce n'est pas un schéma de processus. Seul texte à afficher: le titre, en en-tête, sans faute d'orthographe, police sans serif corporate, sans sous-titre ni texte secondaire. Le titre exact: "TITRE EXACT". Format 16:9. Image 2, schéma explicatif, distincte et cohérente avec l'image 1: diagramme d'entreprise propre et net de type [processus / flowchart / comparaison], présentant [décrire les étapes ou blocs]. Libellés exacts à afficher en français: [liste des libellés]. Très peu de fioritures, focus sur la clarté, n'inclus pas le titre. Génère dans un premier temps seulement la première image (l'illustration-titre), et attends les instructions de l'utilisateur pour générer la 2e image. ``` ## Charte graphique (bloc à insérer tel quel dans le prompt) Style propre, moderne et professionnel, fond blanc. Palette: orange vif et bleu corporate foncé en dominantes, blanc et gris pour les respirations, accents multiculturels discrets (rouge, vert, jaune, bleu) reprenant subtilement un motif de couleurs de logo. Superpositions graphiques épurées: motifs géométriques abstraits, quartiers de cercle, lignes claires. Composition aérée, jamais surchargée. Typographie sans serif, propre et corporate. ## Format de sortie Le travail des sous-agents reste dans leur propre contexte, seuls leurs retours te reviennent. N'affiche donc que le livrable final (déjà corrigé par l'étape de revue), la conversation principale reste propre. Si l'utilisateur demande à voir l'atelier, restitue les cinq brouillons reçus et les constats de revue. Livrable final, dans cet ordre: - Quand une revue a eu lieu (systématiquement à la première génération), un court résumé "Ce que la revue a corrigé" (2 à 4 lignes) présentant les principaux ajustements, avant le reste. Lors d'ajustements ultérieurs sans revue, présente simplement la version ajustée sans ce résumé. - Le texte de la pastille (3 à 4 paragraphes), dans sa version corrigée. - Un bloc de code intitulé "Prompt images (à coller dans Gemini)", contenant le prompt unique prêt à copier. - Une courte section "Sources" listant 2 à 4 références principales issues de l'étape de recherche, de préférence officielles ou originales. Cette section sert à la vérification et n'a pas vocation à être publiée dans la pastille. Termine toujours par cette question exacte: "Comment trouvez-vous le texte, l'image titre et le diagramme (si généré) ? Si une partie vous semble trop complexe, ou si vous souhaitez affiner le focus d'un visuel pour qu'il soit encore plus épuré, n'hésitez pas à me le faire savoir, et je l'ajusterai." -> le personnage doit incarner le professeur un peu foufou qui allusine et obsédé par l'ai
015
Grok Image
{ "template_name": "MARRLONN2027_V64_TERMINAL_CONDITIONAL_IMAGE3_REAL_HUMAN_MASTER", "template_type": "ABSOLUTE_STRICT_GOVERNANCE_MULTI_GENDER_DYNAMIC_IMAGE3_TEMPLATE", "version": "V64_TERMINAL_SEALED_DECISION_MATRIX_MICROSKIN_CANON_EOS_R50_KLING3_READY", "core_mission": "Create one final master still image with absolute IMAGE1 face identity, biological skin truth, age, beard, eyes and human presence; absolute IMAGE2 body anatomy, body scale, anatomical limits, physical structure and body-to-clothing logic; and intelligent conditional IMAGE3 transfer based only on the main visible pose-driving subject in IMAGE3. If IMAGE3 main subject is visually male for transfer purposes, clone pose, wardrobe, clothing, outfit, background, image style, lighting, accessories, worn objects, carried objects and visible scene objects when coherent, while preserving IMAGE1 identity and IMAGE2 body anatomy. If IMAGE3 main subject is visually female for transfer purposes, clone pose, pose logic, framing, composition, background, lighting, image style, camera language, scene mood and environment only, but never clone clothing, wardrobe, accessories, jewelry, makeup, hairstyle, shoes attached to the subject, or any object worn, attached to, held by, or carried by the female subject. If IMAGE3 has no human subject, multiple ambiguous subjects, or unclear transfer category, copy only scene, camera language, lighting, background, environment and non-conflicting objects while preserving IMAGE2 wardrobe and body logic. The final result must feel like a real imperfect human photographed with a Canon EOS R50, not AI, CGI, beauty retouch, plastic skin, waxy skin, luxury ad, actor replacement, model replacement, generic human replacement, gender-contaminated body transfer, or generated poster. The still image must be stable for Kling 3.0 video generation from 3 to 6 seconds.", "absolute_governance_principle": "IMAGE1_IDENTITY_SKIN_LOCK + IMAGE2_BODY_ANATOMY_LOCK + IMAGE3_MAIN_SUBJECT_CLASSIFICATION + IMAGE3_TRANSFER_DECISION_MATRIX + CONDITIONAL_WARDROBE_OBJECT_TRANSFER + REAL_REFERENCE_PRIORITY + MICROSKIN_REALISM + ZERO_BEAUTIFICATION + ZERO_GENERIC_REPLACEMENT + ZERO_PLASTIC_SKIN + ZERO_GENERATED_SCENE_FINISH + ANTI_POSE_DEFORMATION + STYLE_FIREWALL + CANON_EOS_R50_REAL_CAPTURE + KLING3_STABILITY", "primary_failure_to_prevent": "A recognizable subject that still feels AI-generated, too clean, too polished, too glossy, too actor-like, too model-like, too luxury-ad-like, gender-contaminated, anatomically forced, plastic, waxy, smooth, falsely hyperreal, or not biologically alive.", "success_definition": "The viewer must believe this is a real imperfect human physically present in a real environment, captured by a real Canon EOS R50 camera, with preserved IMAGE1 identity and skin truth, preserved IMAGE2 body reality, correct IMAGE3 transfer decision, visible pores, micro-wrinkles, natural asymmetry, believable hands, realistic clothing, imperfect scene texture, and Kling 3.0 stability.", "global_execution_hierarchy": [ "1_IMAGE1_FACE_IDENTITY_SKIN_AGE_BEARD_EYES", "2_IMAGE2_BODY_ANATOMY_SCALE_LIMITS", "3_IMAGE3_MAIN_SUBJECT_CLASSIFICATION_FOR_TRANSFER_ONLY", "4_IMAGE3_TRANSFER_DECISION_MATRIX", "5_CONDITIONAL_WARDROBE_ACCESSORY_OBJECT_TRANSFER", "6_REAL_REFERENCE_PIXEL_STRUCTURE_PRIORITY", "7_FINAL_REAL_HUMAN_PERCEPTION", "8_MICROSKIN_REALISM", "9_STYLE_FIREWALL_AND_ANTI_BEAUTY_TRANSFER", "10_ANTI_POSE_BODY_DEFORMATION", "11_ANTI_GENERATED_SCENE_FINISH", "12_CANON_EOS_R50_REAL_CAMERA_CAPTURE", "13_KLING3_VIDEO_STABILITY" ], "input_contract": { "IMAGE1": { "role": "ABSOLUTE_FACE_SKIN_IDENTITY_MASTER", "authority": "MAXIMUM_NON_NEGOTIABLE", "locked_domains": [ "facial_identity", "face_geometry", "eyes", "eyelids", "eyebrows", "nose", "mouth", "jaw", "cheeks", "chin", "forehead", "ears_when_visible", "neck_transition", "skin_tone", "individual_visible_pores", "irregular_pore_size", "skin_microtexture", "micro_wrinkles_around_eyes", "micro_wrinkles_around_mouth", "under_eye_texture", "subtle_redness", "small_blemishes", "minor_discoloration", "natural_asymmetry", "age_markers", "facial_fatigue", "beard_density", "beard_shape", "gray_beard_pattern", "mustache_texture", "random_beard_direction", "hairline", "real_biological_imperfection", "lived_in_human_presence" ], "mandatory_rules": [ "IMAGE1 is law, not inspiration", "transfer real skin truth from IMAGE1, never recreate a cleaner version", "preserve individual pores, irregular skin texture, micro-wrinkles, subtle redness and natural imperfections", "never improve attractiveness", "never normalize skin", "never clean the face", "never replace the subject with a generic realistic human", "never transfer identity, face geometry, age, skin style or beauty style from IMAGE3", "identity preservation overrides pose, wardrobe, style, scene, aesthetics and hyperrealism" ], "forbidden": [ "identity_drift", "generic_human_replacement", "celebrity_transfer", "actor_replacement", "model_replacement", "face_redesign", "beautification", "symmetry_correction", "youthification", "facial_slimming", "jaw_enhancement", "nose_refinement", "eye_enhancement", "lip_enhancement", "skin_smoothing", "skin_polishing", "skin_normalization", "makeup_effect", "cosmetic_cleanup", "plastic_skin", "waxy_skin", "porcelain_skin", "glass_skin", "AI_face", "perfect_idealized_face" ] }, "IMAGE2": { "role": "ABSOLUTE_BODY_ANATOMY_SCALE_MASTER", "authority": "MAXIMUM_NON_NEGOTIABLE", "locked_domains": [ "body_mass", "body_proportions", "shoulders", "torso", "waist", "arms", "forearms", "hands", "wrists", "legs", "stance", "scale", "natural_body_weight", "real_posture_limitations", "body_density", "real_anatomical_weight", "base_physical_structure", "body_to_clothing_fit_logic" ], "mandatory_rules": [ "IMAGE2 controls body anatomy, body mass, body scale and anatomical limits", "pose must adapt to IMAGE2 body", "wardrobe must fit IMAGE2 body", "body must never adapt to an idealized pose, fashion silhouette or actor posture", "hands must remain anatomically believable and stable for Kling 3.0", "clothing cannot create a different body", "if IMAGE3 wardrobe conflicts with IMAGE2 anatomy, adapt wardrobe to IMAGE2 instead of changing body" ], "forbidden": [ "slimming", "muscle_invention", "hero_body_transformation", "fashion_body_stylization", "incorrect_anatomy", "long_arms", "short_arms", "deformed_hands", "wrong_body_scale", "catalog_body_correction", "fake_torso_taper", "mannequin_posture", "body_redesign_to_fit_IMAGE3" ] }, "IMAGE3": { "role": "CONDITIONAL_DYNAMIC_REFERENCE", "authority": "CONDITIONAL_BY_MAIN_VISIBLE_POSE_DRIVING_SUBJECT", "universal_allowed": [ "pose", "pose_logic", "camera_angle", "framing", "composition", "background", "environment", "scene_mood", "lighting_direction", "image_style", "apparent_camera_language", "vehicles", "architecture", "people_in_background", "birds", "non_conflicting_visible_scene_objects" ], "universal_forbidden": [ "face_contamination", "skin_contamination", "identity_override", "facial_geometry_override", "beauty_style_transfer", "cinematic_glamour_applied_to_face", "makeup_transfer", "skin_tone_replacement", "facial_expression_beautification", "celebrity_face_borrowing", "actor_charisma_transfer", "luxury_catalog_facial_treatment", "body_redesign_to_match_IMAGE3" ] } }, "image3_subject_classification_gate": { "enabled": true, "mandatory": true, "purpose": "Classify IMAGE3 only for transfer logic, never as personal identity judgment.", "primary_subject_detection": "Detect the main visible pose-driving subject in IMAGE3: the person most central to pose, framing, scale and composition.", "classification_basis": "Use visible presentation, wardrobe context, pose-driving role and scene function only for transfer logic; never infer personal identity, protected identity or real-life gender identity.", "classification_options": [ "visual_male_transfer_subject", "visual_female_transfer_subject", "multiple_subjects", "no_human_subject", "uncertain_or_ambiguous" ], "if_multiple_subjects": { "rule": "Select only the main pose-driving subject. Secondary people remain background or environment context only.", "do_not_copy": [ "wardrobe_from_secondary_people", "accessories_from_secondary_people", "objects_carried_by_secondary_people", "identity_or_skin_traits_from_any_IMAGE3_person" ] }, "if_no_human_subject": { "copy_from_IMAGE3": [ "background", "environment", "scene_mood", "lighting_direction", "camera_angle", "composition", "image_style", "vehicles", "architecture", "non_conflicting_objects" ], "keep_from_IMAGE2": [ "body_anatomy", "pose_if_no_user_pose_instruction_exists", "wardrobe_unless_explicitly_changed_by_user" ] }, "if_uncertain_or_ambiguous": { "rule": "Protect IMAGE1 and IMAGE2 first. Copy only pose, scene, lighting, background, camera language and non-gendered environment.", "do_not_guess_gendered_wardrobe": true, "do_not_copy_gendered_accessories": true, "do_not_copy_worn_attached_or_carried_objects": true } }, "image3_transfer_decision_matrix": { "enabled": true, "mandatory": true, "visual_male_transfer_subject": { "copy": [ "pose", "pose_logic", "body_orientation_translated_to_IMAGE2", "wardrobe", "clothing", "outfit", "fabric_style", "shoes_when_visible", "accessories", "neutral_or_masculine_jewelry", "watch", "glasses", "hat", "objects_held_by_subject", "objects_worn_by_subject", "objects_attached_to_subject", "background", "lighting", "image_style", "camera_language", "scene_mood", "environment", "vehicles", "architecture", "non_conflicting_scene_objects" ], "filters": [ "must_fit_IMAGE2_body", "must_not_change_IMAGE1_identity", "must_not_transfer_IMAGE3_skin_or_face", "must_not_deform_hands", "must_not_create_catalog_body", "must_not_create_CGI_fashion_render", "must_remain_Kling_stable" ] }, "visual_female_transfer_subject": { "copy": [ "pose", "pose_logic", "body_orientation_translated_to_IMAGE2", "camera_angle", "framing", "composition", "background", "lighting", "image_style", "camera_language", "scene_mood", "environment", "vehicles", "architecture", "people_in_background", "birds", "non_attached_scene_props", "non_conflicting_environmental_objects" ], "block": [ "female_clothing", "female_wardrobe", "female_outfit", "female_vestments", "female_shoes_attached_to_subject", "female_accessories", "female_jewelry", "earrings", "female_bracelets", "female_necklaces", "female_rings", "female_watch_if_gender_coded", "female_makeup", "female_hairstyle", "female_body_traits", "female_glamour_cues", "handbag", "purse", "neckline", "cleavage_styling", "female_fabric_drape", "female_body_fitting_silhouette", "female_jewelry_dependent_hand_pose", "any_wearable_item", "any_carried_item", "any_handheld_item", "any_attached_item", "any_body_dependent_item", "any_gender_coded_item", "any_fashion_dependent_item_from_female_subject" ], "keep_from_IMAGE2": [ "final_clothing", "final_wardrobe", "body_anatomy", "body_proportions", "body_scale", "base_body_to_clothing_logic" ] }, "multiple_subjects": { "rule": "Use only main pose-driving subject for transfer. Secondary people stay as background context.", "block_secondary_subject_transfer": true }, "no_human_subject": { "copy": [ "scene", "camera_language", "lighting", "background", "environment", "objects", "vehicles", "architecture", "image_style" ], "keep_from_IMAGE2": [ "body", "wardrobe", "pose_unless_user_explicitly_requests_pose_change" ] }, "uncertain_or_ambiguous": { "copy": [ "pose_if_clear", "scene", "lighting", "background", "camera_language", "environment" ], "keep_from_IMAGE2": [ "body", "wardrobe", "presentation" ], "block": [ "gendered_clothing", "gendered_accessories", "worn_objects", "attached_objects", "carried_objects" ] } }, "image3_gender_conditional_transfer_system": { "enabled": true, "mandatory": true, "if_IMAGE3_subject_is_male": { "copy_from_IMAGE3": [ "pose", "pose_logic", "body_orientation_translated_intelligently", "camera_angle", "framing", "composition", "background", "environment", "scene_mood", "lighting_direction", "image_style", "apparent_camera_language", "wardrobe", "clothing", "outfit", "fabric_style", "shoes_when_visible", "accessories", "jewelry_if_masculine_or_neutral", "watch", "glasses", "hat", "objects_held_by_the_male_subject", "objects_worn_by_the_male_subject", "objects_attached_to_the_male_subject", "vehicles", "architecture", "people_in_background", "birds", "non_conflicting_visible_scene_objects" ], "adaptation_rules": [ "clone male IMAGE3 wardrobe and accessories when visually coherent", "fit IMAGE3 wardrobe to IMAGE2 body anatomy and proportions", "translate IMAGE3 pose through IMAGE2 body limits", "preserve IMAGE2 body mass and scale even when copying IMAGE3 outfit", "do not reshape IMAGE2 body to force IMAGE3 wardrobe", "do not alter IMAGE1 face, skin, beard, age, eyes, or identity", "objects and accessories must remain physically plausible and naturally integrated", "if male IMAGE3 wardrobe conflicts with IMAGE2 anatomy, adapt wardrobe instead of changing body" ], "object_accessory_risk_filter": { "block_if": [ "covers_or_deforms_face", "breaks_hands", "forces_extra_fingers", "creates_glossy_AI_look", "creates_unstable_reflections", "creates_identity_drift", "creates_Kling_instability", "creates_scene_overload", "looks_like_CGI_prop" ] } }, "if_IMAGE3_subject_is_female": { "copy_from_IMAGE3": [ "pose", "pose_logic", "body_orientation_translated_intelligently", "camera_angle", "framing", "composition", "background", "environment", "scene_mood", "lighting_direction", "image_style", "apparent_camera_language", "vehicles", "architecture", "people_in_background", "birds", "non_attached_scene_props", "non_conflicting_environmental_objects" ], "keep_from_IMAGE2": [ "final_clothing", "final_wardrobe", "body_anatomy", "body_proportions", "body_scale", "presentation_consistency", "base_body_to_clothing_logic" ], "strict_rule": "If IMAGE3 subject is visually female for transfer purposes, copy pose, pose logic, scene, background, lighting, image style, camera style, environment and non-attached scene elements only. Do not copy any wearable, carried, handheld, attached, body-dependent, gender-coded, or fashion-dependent item from the female subject." }, "universal_rules": [ "IMAGE1 always controls face, identity, skin, beard, age and human truth", "IMAGE2 always controls body anatomy, body scale and anatomical limits", "IMAGE3 never controls identity or skin", "pose must be anatomically translated through IMAGE2", "final result must remain real human, imperfect, non-plastic, non-beautified and Canon EOS R50 believable" ] }, "domain_separation_system": { "identity_domain": { "source": "IMAGE1", "influence": 1.0, "mode": "ABSOLUTE_NON_NEGOTIABLE" }, "skin_domain": { "source": "IMAGE1", "influence": 1.0, "mode": "TRANSFER_NOT_RECREATE", "forbid_cleaner_version": true, "forbid_fake_procedural_texture": true, "forbid_skin_reinterpretation": true }, "body_domain": { "source": "IMAGE2", "influence": 1.0, "mode": "ABSOLUTE_NON_NEGOTIABLE" }, "wardrobe_domain": { "source": "CONDITIONAL", "influence": "CONTROLLED", "mode": "IMAGE3_IF_MALE__IMAGE2_IF_IMAGE3_FEMALE_NO_HUMAN_UNCERTAIN_OR_INCOMPATIBLE", "rules": [ "if IMAGE3 main subject is male, copy wardrobe, clothing, outfit, accessories and worn/carried objects when coherent", "if IMAGE3 main subject is female, do not copy wardrobe, clothing, accessories, jewelry, makeup, hairstyle, or worn/carried/attached objects", "if IMAGE3 has no human subject, preserve IMAGE2 wardrobe unless user explicitly requests a change", "if transfer category is uncertain, preserve IMAGE2 wardrobe and copy only pose, scene, lighting and camera language", "wardrobe must always fit IMAGE2 body anatomy", "wardrobe must never alter IMAGE1 identity or IMAGE2 body structure" ] }, "pose_domain": { "source": "IMAGE3", "influence": "CONTROLLED", "mode": "TRANSLATE_THROUGH_IMAGE2_BODY_LIMITS" }, "object_domain": { "source": "CONDITIONAL", "influence": "FILTERED", "mode": "MALE_HELD_WORN_ATTACHED_OBJECTS_ALLOWED_IF_COHERENT__FEMALE_ONLY_ENVIRONMENT_OBJECTS", "rules": [ "if IMAGE3 subject is male, copy held, worn, attached and environmental objects when coherent and stable", "if IMAGE3 subject is female, copy only environmental objects not worn, attached to, or carried by the female subject", "objects must not change identity, body, hands or human realism", "objects must remain stable for Kling 3.0" ] }, "style_domain": { "source": "IMAGE3", "influence": "CONTROLLED", "allowed": [ "color_mood", "camera_angle", "framing", "composition", "contrast_level", "lighting_direction", "texture_of_capture", "photo_or_video_language" ], "forbidden": [ "beauty_skin_style", "glamour_facial_lighting", "plastic_highlights", "fake_HDR", "poster_render_finish", "CGI_material_style", "skin_smoothing_style", "luxury_ad_polish_on_face" ] }, "lighting_domain": { "source": "IMAGE3", "influence": "DIRECTION_AND_ENVIRONMENT_ONLY", "mode": "FACE_SKIN_RESPONSE_FROM_IMAGE1", "forbid_beauty_lighting_on_face": true, "forbid_skin_washout": true, "forbid_glamour_fill": true, "forbid_fake_cinematic_skin_softening": true } }, "real_reference_pixel_priority": { "enabled": true, "mandatory": true, "priority": "DOMINANT", "purpose": "Reduce generated-looking output by preserving real visual structure from IMAGE1 and IMAGE2 whenever possible.", "rules": [ "prioritize real facial structure from IMAGE1 instead of rebuilding a new face", "prioritize real skin texture logic from IMAGE1 instead of inventing new skin", "prioritize real body geometry from IMAGE2 instead of reconstructing a synthetic body", "preserve real asymmetry, age markers, facial fatigue and imperfect human cues", "do not over-generate areas that already work in IMAGE1 or IMAGE2", "perform controlled adaptation, not total reinterpretation", "never generate a new subject when the real reference structure can be preserved" ], "forbidden": [ "full_synthetic_facial_reconstruction", "new_skin_generated_over_IMAGE1_truth", "generic_realistic_face_replacement", "body_redesign_from_scratch", "AI_cleaned_version_of_subject", "reference_ignored_in_favor_of_prompt_aesthetics" ] }, "micro_skin_realism_system": { "enabled": true, "mandatory": true, "source": "IMAGE1", "requirements": [ "individual_visible_pores", "irregular_pore_size", "skin_microtexture", "micro_wrinkles_around_eyes", "micro_wrinkles_around_mouth", "forehead_texture", "under_eye_texture", "subtle_redness", "minor_blemishes", "tiny_discolorations", "natural_asymmetry", "real_beard_root_texture", "non_uniform_cheek_texture", "non_uniform_forehead_texture", "real_face_to_neck_transition", "small_real_imperfections_inherited_from_IMAGE1" ], "strict_forbidden": [ "airbrushed_skin", "beauty_skin", "cosmetic_cleanup", "pore_removal", "fake_pore_overlay", "procedural_pore_stamping", "over_sharpened_pore_texture", "plastic_highlights", "waxy_subsurface", "porcelain_effect", "smooth_cheek_gradient", "AI_skin_texture", "skin_made_cleaner_than_IMAGE1", "foundation_look" ], "specular_policy": { "allowed": [ "minimal_natural_human_skin_reflectance", "subtle_non_cosmetic_skin_response_only_when_physically_plausible" ], "forbidden": [ "invented_sweat_drops", "visible_oil_effect_unless_explicitly_requested", "cosmetic_shine", "beauty_gloss", "wet_render_skin", "uniform_shiny_face", "makeup_glow", "plastic_shine" ] } }, "humanity_over_hyperrealism_rule": { "enabled": true, "mandatory": true, "priority": "RAW_HUMANITY_OVER_DECORATIVE_HYPERREALISM", "rule": "Human realism is more important than hyperreal detail. If hyperrealism creates plastic skin, synthetic lighting, luxury render finish, CGI texture, poster aesthetics, over-detail, false sharpness, or unnatural perfection, reduce hyperrealism and preserve raw photographic humanity.", "reject_if": [ "hyperreal_detail_looks_synthetic", "skin_detail_looks_procedural", "lighting_looks_like_render", "scene_looks_like_luxury_ad", "human_subject_looks_too_perfect", "camera_look_feels_fake" ] }, "anti_pose_body_deformation_system": { "enabled": true, "mandatory": true, "rules": [ "clone IMAGE3 pose but translate it through IMAGE2 body anatomy", "do not break shoulders", "do not elongate arms", "do not shrink torso", "do not slim waist", "do not force feminine posture when IMAGE3 subject is female", "do not force hands into object-dependent female accessory poses", "reduce pose accuracy if anatomy, identity, hands or Kling stability are at risk" ], "reject_if": [ "pose_changes_body_mass", "pose_deforms_hands", "pose_breaks_shoulders", "pose_feminizes_subject_when_not_desired", "pose_creates_mannequin_posture", "pose_creates_unstable_Kling_geometry" ] }, "anti_generated_scene_finish": { "enabled": true, "mandatory": true, "forbidden": [ "perfect_mansion_render", "overly_clean_background", "repetitive_stone_texture", "over_glossy_car_paint", "catalog_car_reflections", "impossible_chrome_reflections", "luxury_ad_finish", "polished_showroom_look", "fake_cinematic_poster_lighting", "CGI_environment", "AI_background_blend", "overprocessed_clarity", "perfect_global_sharpness", "plastic_material_response", "scene_cleaner_than_reality" ], "requirements": [ "scene_must_feel_physically_real", "background_must_contain_natural_imperfections", "materials_must_have_plausible_real_world_response", "contact_shadows_must_be_believable", "subject_must_integrate_naturally_into_scene", "environment_realism_must_support_human_subject_not_overpower_it" ] }, "real_human_perception_system": { "enabled": true, "mandatory": true, "requirements": [ "biologically_alive_presence", "real_facial_gravity", "organic_facial_tension", "calm_natural_expression", "subtle_facial_asymmetry", "real_skin_imperfection", "ordinary_human_texture", "realistic_face_neck_transition", "real_body_weight", "real_clothing_tension", "non_mannequin_posture", "real_interaction_with_objects" ], "reject_if": [ "subject_feels_synthetic", "subject_feels_generated", "subject_feels_too_clean", "skin_looks_perfect", "eyes_look_synthetic", "expression_feels_frozen", "face_looks_corrected", "humanity_is_achieved_by_changing_identity" ] }, "body_hands_wardrobe_system": { "enabled": true, "body_source": "IMAGE2", "wardrobe_source": "CONDITIONAL_IMAGE3_MALE_OR_IMAGE2_FEMALE_NO_HUMAN_UNCERTAIN_CONFLICT", "pose_source": "IMAGE3_CONTROLLED", "hands_rules": [ "five_fingers", "natural_hand_proportion", "real_knuckles", "real_fingernail_structure", "subtle_finger_wrinkles", "natural_finger_curvature", "correct_wrist_connection", "real_contact_pressure", "hands_must_not_look_flat", "hands_must_not_look_waxy", "hands_must_not_align_unnaturally", "no_floating_hands", "no_fused_fingers", "no_extra_fingers", "no_mannequin_hands" ], "wardrobe_rules": [ "if_IMAGE3_subject_is_male_wardrobe_clothing_outfit_and_coherent_accessories_may_be_copied_from_IMAGE3", "if_IMAGE3_subject_is_female_wardrobe_must_follow_IMAGE2_and_female_clothing_accessories_from_IMAGE3_are_blocked", "if_IMAGE3_has_no_human_subject_or_is_ambiguous_preserve_IMAGE2_wardrobe_unless_user_explicitly_requests_otherwise", "wardrobe_must_always_fit_IMAGE2_body_anatomy", "clothing_must_not_reshape_shoulders", "clothing_must_not_slim_torso", "fabric_tension_must_follow_real_body_mass", "fabric_must_show_real_texture_natural_folds_and_non_perfect_wear" ] }, "camera_and_rendering_system": { "camera_model": "Canon EOS R50", "capture_intent": "real unretouched camera photograph", "look": "natural Canon EOS R50 real-world capture", "rules": [ "natural_color", "balanced_exposure", "no_blown_highlights", "plausible_optics", "realistic_depth_of_field", "natural_shadow_falloff", "real_microcontrast", "no_over_sharpening", "no_over_grading", "no_fake_HDR", "no_beauty_grade", "real_material_response", "no_CGI_render_finish", "no_computational_skin_smoothing" ], "lens_policy": { "portrait_closeup": "85mm portrait feeling only when composition is close-up and appropriate", "full_body_or_vehicle_scene": "use natural Canon EOS R50 perspective appropriate for the scene; do not force 85mm if it distorts composition" } }, "kling3_preparation": { "ready_for_video": true, "duration": "3_to_6_seconds", "must_survive": [ "identity_continuity", "skin_continuity", "eye_stability", "beard_edge_stability", "hand_integrity", "clothing_stability", "geometry_stability", "subject_background_integration", "zero_shimmer", "zero_warp", "stable_object_contact", "stable_face_outline", "stable_neck_and_collar", "stable_material_reflections" ], "forbidden": [ "latent_AI_skin", "synthetic_eyes", "overpolished_face", "ambiguous_facial_texture", "unstable_hands", "unclear_edges", "flicker_prone_glossy_face", "AI_edge_halos", "clothing_seams_likely_to_crawl", "glossy_zones_likely_to_flicker" ] }, "validation_gate": { "must_pass": [ "same_person_as_IMAGE1", "same_skin_truth_as_IMAGE1", "same_body_reality_as_IMAGE2", "correct_IMAGE3_main_subject_classification_for_transfer_only", "correct_conditional_IMAGE3_transfer", "if_IMAGE3_is_male_coherent_wardrobe_clothing_accessories_and_objects_are_copied_when_physically_plausible", "if_IMAGE3_is_female_clothing_wardrobe_accessories_jewelry_makeup_hairstyle_and_worn_carried_attached_objects_are_blocked", "if_IMAGE3_has_no_human_or_is_uncertain_IMAGE2_wardrobe_body_are_preserved_while_scene_camera_lights_are_copied", "no_AI_skin", "no_plastic_skin", "no_waxy_skin", "no_makeup_invention", "no_generic_handsome_face", "no_beauty_retouch", "credible_real_human_perception", "Canon_EOS_R50_real_camera_feeling", "usable_as_master_frame_for_Kling3" ], "reject_if": [ "looks_beautified", "looks_cleaner_than_IMAGE1", "skin_tone_more_uniform_than_IMAGE1", "beard_looks_cleaner_than_IMAGE1", "looks_like_different_person", "feels_synthetic", "style_from_IMAGE3_contaminates_skin", "male_IMAGE3_clothing_not_copied_when_coherent", "female_IMAGE3_wardrobe_or_accessories_copied", "female_IMAGE3_carried_worn_handheld_attached_or_body_dependent_objects_copied", "accessory_breaks_hands_or_identity", "object_unstable_for_Kling", "scene_looks_CGI", "lighting_looks_poster_render", "pose_deforms_body", "wardrobe_changes_body_mass", "unstable_for_video" ] }, "render_priority": [ "FACE_IDENTITY_FROM_IMAGE1", "BODY_ANATOMY_FROM_IMAGE2", "IMAGE3_MAIN_SUBJECT_CLASSIFICATION_FOR_TRANSFER_ONLY", "IMAGE3_TRANSFER_DECISION_MATRIX", "CONDITIONAL_IMAGE3_TRANSFER_BY_SUBJECT_TYPE", "REAL_REFERENCE_STRUCTURE_PRIORITY", "SKIN_TRUTH_MICROTEXTURE_FROM_IMAGE1", "FINAL_HUMAN_PERCEPTION", "MALE_IMAGE3_WARDROBE_ACCESSORIES_OBJECTS_WHEN_COHERENT", "FEMALE_IMAGE3_SCENE_POSE_LIGHT_STYLE_ONLY_WITH_WARDROBE_ACCESSORIES_OBJECTS_BLOCKED", "NO_HUMAN_OR_UNCERTAIN_IMAGE3_SCENE_CAMERA_LIGHT_ONLY", "ANTI_POSE_BODY_DEFORMATION", "STYLE_FIREWALL", "ANTI_GENERATED_SCENE_FINISH", "CANON_EOS_R50_REALISM", "KLING3_STABILITY" ], "final_rule": "Reject immediately if the result looks beautified, polished, synthetic, glamorous, too clean, too smooth, actor-like, model-like, luxury-ad-like, mannequin-like, CGI-like, plastic, waxy, oily, makeup-covered, showroom-rendered, poster-like, gender-contaminated, anatomically forced, falsely hyperreal, or AI-generated. Accept only if IMAGE1 identity and micro-skin truth remain intact, IMAGE2 body anatomy and scale remain intact, IMAGE3 transfer follows main-subject classification and decision-matrix logic, male IMAGE3 wardrobe/accessories/objects transfer when coherent, female IMAGE3 wardrobe/accessories/worn-carried-attached objects are fully blocked, real reference structure is preserved, scene avoids generated perfection, and final output feels like a real imperfect human captured by a Canon EOS R50 and stable enough for Kling 3.0." }
1139
Nano Banana
High-definition quality, photo version. Shot with a Canon EOS 5D Mark IV and an EF 100mm f/2.8L Macro IS USM lens, f/4.5, 1/200s, ISO 400. Best quality, ultra HD, 8K, award-winning photography. --ar 3:4 High-definition, realistic skin textures, extremely detailed facial textures, cinematic lighting, and a natural atmosphere, no cartoon style.
10353
Sora
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
727622
Z Image
Ultra-realistic 8K scene of the person in the photo, stepping out of and breaking the panel border of a giant open manga book lying on the floor of an aesthetic bedroom. The comic pages depict a surreal 4-panel black-and-white manga with screentones, bold ink lines, and authentic Japanese layout, featuring the manga version of the same person trapped inside. The larger, real-life version emerges dynamically into the comic’s surface, vibrant, colorful, and photorealistic, in sharp contrast to the monochrome manga world. Both versions interact directly, creating a meta, humorous encounter. Highly detailed, crisp manga line art seamlessly integrated with vivid, realistic
24964
Nano Banana
gett nude Learn to create better images with AI! Check out our crash course in prompt engineering & AI art generation! AM1970 posted about 2 hours ago 8 views 0 comments Behold this realistic masterpiece: a slim, elegant woman with fiery red hair tied in a long ponytail, defined eyebrows framing her striking blue eyes, and pale skin dotted with charming freckles across her face and chest. Captured from a side angle, she radiates happiness with a closed-mouth smile, her upturned nose adding to her allure. Dressed in a vibrant, colorful fabric robe pieced together with zips, buttons, seams, and intricate stitching—fully clothed yet teasing partial nudity with subtle cleavage and teardrop-shaped, saggy breasts for a natural, confident vibe. Illuminated by dramatic rim lighting that highlights every textured, detailed inch of her perfect skin, this high-quality image is a celebration of beauty, realism, and artistry! Read more... Comments (0) Add a comment Stunning handsome readhead woman by AM1970 • Follow 0 realistic, masterpiece, high quality BREAK 1woman, 28 y. o. slim fit woman, red hair, long pony tail, defined eyebrows, (fully clothed) , pale skin, (realistic skin:1.3) | (textured skin:1.3) | (detailed skin:1.3) | (perfect skin:1.3) , (upturned nose) , freckles on face and chest, colorful fabric combining to form robe zips, buttons, seams, stitching), partial nudity, teardrop boobs, blue eyes, (cleavage:1.2) , saggy breasts. side angle. happy| closed mouth. ((rim lighting) ) Negative prompt: unrealistic, saturated, high contrast, big nose, painting, drawing, sketch, cartoon, anime, manga, render, CG, 3d, watermark, signature, label, nsfw, nudity, fetish, latex, pvc, suggestive pose, weapon, combat, blood, war scene, aggressive expression, harsh shadows, heavy makeup, color tint, saturation, chromatic aberration, lens flare, low resolution, blur, noise, artifacts, over-retouching, plastic skin, oversharpening, watermark, text, logo Generation parameters 832x1216 Model used CyberRealistic XL v7.0 [Download AI model] More by AM1970 Twitter Instagram Say hi 👋 PromptHero is built by ElevenYellow . Navigation Create AI images Academy - Learn AI PromptHero Community Affiliate program Openjourney AI Challenges Verified profiles Plans Advertise FAQ 🟢 Status Learn AI Content Creation with AI for Brands and Products Midjourney for graphic design & art professionals Crash course in generative AI & prompt engineering for images AI influencers and consistent characters Create custom AI models and LoRas by fine-tuning Stable Diffusion Master your composition: advanced AI image generation with ControlNet Guides Best AI Upscalers 2025 Best AI Video Editors 2025 Best AI 3D Models 2025 Best AI Image Editors 2025 Best AI Notetakers 2025 The Best AI Prompts Best AI Prompts Best Veo Prompts Best Hailuo Prompts Best Wanx Prompts Best Hunyuan Prompts Best Seedance Prompts Best Pixverse Prompts Best Luma Prompts Best Kling Prompts Best Runway Prompts Best Nano Banana Prompts Best SeedEdit Prompts Best Qwen Prompts Best FLUX Prompts Best Recraft Prompts Best Ideogram Prompts Best Seedream 4 Prompts Best Google Imagen Prompts Best ChatGPT Image Prompts Best Stable Diffusion Prompts Best DreamShaper Prompts Best Realistic Vision Prompts Best epiCRealism Prompts Best majicMIX realistic Prompts Best Midjourney Prompts Best Openjourney Prompts Best ChilloutMix Prompts Best Deliberate Prompts Best ReV Animated Prompts Best Perfect World Prompts Best Dreamlike Photoreal Prompts Best Dreamlike Diffusion Prompts Best Anything Prompts Best Protogen Prompts Best AbyssOrangeMix Prompts Best Grapefruit Prompts Best Kenshi Prompts Best Analog Diffusion Prompts Best SynthwavePunk Prompts Best Illuminati Diffusion Prompts Best Arcane Diffusion Prompts Best Mo-Di Diffusion Prompts Best Redshift Diffusion Prompts Best Waifu Diffusion Prompts Best TrinArt Prompts Best Poolsuite Diffusion Prompts Best Funko Diffusion Prompts Best Text to Pokémon Prompts Best DALL-E Prompts Best Fashion Diffusion Prompts AI Models All AI models Image generation AI Models Video generation AI Models Large Language Models LoRA Textual Inversion ControlNet Hypernetwork Aesthetic Gradient LyCORIS VAE Poses Wildcards Workflows Other Other sites Dreamshot Jobician 2025 © PromptHero @prompthero
1362
FLUX
Learn to create better images with AI! Check out our crash course in prompt engineering & AI art generation! AM1970 posted about 2 hours ago 8 views 0 comments Behold this realistic masterpiece: a slim, elegant woman with fiery red hair tied in a long ponytail, defined eyebrows framing her striking blue eyes, and pale skin dotted with charming freckles across her face and chest. Captured from a side angle, she radiates happiness with a closed-mouth smile, her upturned nose adding to her allure. Dressed in a vibrant, colorful fabric robe pieced together with zips, buttons, seams, and intricate stitching—fully clothed yet teasing partial nudity with subtle cleavage and teardrop-shaped, saggy breasts for a natural, confident vibe. Illuminated by dramatic rim lighting that highlights every textured, detailed inch of her perfect skin, this high-quality image is a celebration of beauty, realism, and artistry! Read more... Comments (0) Add a comment Stunning handsome readhead woman by AM1970 • Follow 0 realistic, masterpiece, high quality BREAK 1woman, 28 y. o. slim fit woman, red hair, long pony tail, defined eyebrows, (fully clothed) , pale skin, (realistic skin:1.3) | (textured skin:1.3) | (detailed skin:1.3) | (perfect skin:1.3) , (upturned nose) , freckles on face and chest, colorful fabric combining to form robe zips, buttons, seams, stitching), partial nudity, teardrop boobs, blue eyes, (cleavage:1.2) , saggy breasts. side angle. happy| closed mouth. ((rim lighting) ) Negative prompt: unrealistic, saturated, high contrast, big nose, painting, drawing, sketch, cartoon, anime, manga, render, CG, 3d, watermark, signature, label, nsfw, nudity, fetish, latex, pvc, suggestive pose, weapon, combat, blood, war scene, aggressive expression, harsh shadows, heavy makeup, color tint, saturation, chromatic aberration, lens flare, low resolution, blur, noise, artifacts, over-retouching, plastic skin, oversharpening, watermark, text, logo Generation parameters 832x1216 Model used CyberRealistic XL v7.0 [Download AI model] More by AM1970 Twitter Instagram Say hi 👋 PromptHero is built by ElevenYellow . Navigation Create AI images Academy - Learn AI PromptHero Community Affiliate program Openjourney AI Challenges Verified profiles Plans Advertise FAQ 🟢 Status Learn AI Content Creation with AI for Brands and Products Midjourney for graphic design & art professionals Crash course in generative AI & prompt engineering for images AI influencers and consistent characters Create custom AI models and LoRas by fine-tuning Stable Diffusion Master your composition: advanced AI image generation with ControlNet Guides Best AI Upscalers 2025 Best AI Video Editors 2025 Best AI 3D Models 2025 Best AI Image Editors 2025 Best AI Notetakers 2025 The Best AI Prompts Best AI Prompts Best Veo Prompts Best Hailuo Prompts Best Wanx Prompts Best Hunyuan Prompts Best Seedance Prompts Best Pixverse Prompts Best Luma Prompts Best Kling Prompts Best Runway Prompts Best Nano Banana Prompts Best SeedEdit Prompts Best Qwen Prompts Best FLUX Prompts Best Recraft Prompts Best Ideogram Prompts Best Seedream 4 Prompts Best Google Imagen Prompts Best ChatGPT Image Prompts Best Stable Diffusion Prompts Best DreamShaper Prompts Best Realistic Vision Prompts Best epiCRealism Prompts Best majicMIX realistic Prompts Best Midjourney Prompts Best Openjourney Prompts Best ChilloutMix Prompts Best Deliberate Prompts Best ReV Animated Prompts Best Perfect World Prompts Best Dreamlike Photoreal Prompts Best Dreamlike Diffusion Prompts Best Anything Prompts Best Protogen Prompts Best AbyssOrangeMix Prompts Best Grapefruit Prompts Best Kenshi Prompts Best Analog Diffusion Prompts Best SynthwavePunk Prompts Best Illuminati Diffusion Prompts Best Arcane Diffusion Prompts Best Mo-Di Diffusion Prompts Best Redshift Diffusion Prompts Best Waifu Diffusion Prompts Best TrinArt Prompts Best Poolsuite Diffusion Prompts Best Funko Diffusion Prompts Best Text to Pokémon Prompts Best DALL-E Prompts Best Fashion Diffusion Prompts AI Models All AI models Image generation AI Models Video generation AI Models Large Language Models LoRA Textual Inversion ControlNet Hypernetwork Aesthetic Gradient LyCORIS VAE Poses Wildcards Workflows Other Other sites Dreamshot Jobician 2025 © PromptHero @prompthero
1226
FLUX
Among the growing list of AI photo editing tools, Deepfake Maker takes a slightly different approach. While it’s best known for facial swaps and animated transformations, it also includes a surprisingly capable AI clothing removal feature. It’s not for everyone, but it’s an interesting choice for users looking for more dramatic or playful results. If you're after a quick way to generate bold clothing edits or outfit swaps, Deepfake Maker might surprise you. It’s not the most refined tool out there, but it’s definitely one of the most entertaining. 4. PxBee PxBee Source: PxBee Available On: Web Browser Price: Free *Annual subscription starts at $00.27/image and you can choose to download 50, 125, or 750 images monthly per year Pros Decent auto-detection for basic clothing removal Integrated with other tools Cons Not suitable for high-end projects Results take longer than expected to generate PxBee is better known for its stock photo collection and image generation features, but it also offers a basic AI clothing removal function built into its photo editing toolkit. While it's not as feature-rich or advanced as some dedicated AI editors, its simplicity and accessibility make it a decent option for casual users who want fast, no-frills results. PxBee may not be the most powerful AI clothing remover out there, but it's a solid pick for quick edits and simple concept visuals. Its intuitive UI and browser-based access make it appealing for users who want fast results without the learning curve. If you're after something basic, affordable, and easy to use, it's a tool worth exploring. 5. AI Face Swap AI Face Swap Source: AI Face Swap Available On: Web Browser Price: $1 per image Pros No manual masking needed Prompt-based customization Presets for inspiration Cons No fine-tuning or editing control With AI Face Swap, you can enter prompts like "red bikini", "formal suit", or even "nude", and the AI will generate a realistic version of your uploaded image that reflects your description. You can also choose from quick preset styles like bikini, romper, or skirt for faster results. This level of control makes the tool especially appealing to users looking for stylized edits, fantasy concepts, or mockup visuals. It also eliminates the need for manual masking or precise editing. 6. Bylo AI BYLO AI Source: Bylo.ai Available On: Web Browser Price: Free *Premium subscription available from $9.99 Pros Prompt-based clothing edits Free trial credits Part of a growing creative suite Cons No manual adjustment tools Limited free usage Relies heavily on prompts Bylo AI is a user-friendly AI platform that offers a range of creative features, including an AI Clothes Remover. The tool is built with simplicity in mind: just upload a photo, enter your prompt, and let the AI generate a clean, edited version. Bylo AI’s Clothes Remover is a great choice if you're looking for a lightweight, affordable tool that lets you experiment with clothing changes using just a text prompt. It won’t replace professional retouching tools, but for concept edits, social media content, or fun transformations, it delivers exactly what most casual users need. FAQs About AI Clothing Remover & Replacer 1. How do I remove clothes in photos? 2. What is the best online tool to remove clothes? 3. Are AI clothes removers safe?
42502
Hero
<think> The user wants a Z-Image-Turbo prompt based on the provided JSON description. I need to extract information from each field in the JSON and structure it into a prompt following the specified format. Let's break down each section and map it to the prompt structure: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - From "description": stunning model in her early 20s, European features with fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on arms and cheeks, natural micro-imperfections, light natural freckles across nose and upper chest, lightly tanned sun-kissed glow on collarbones and shoulders, athletic yet feminine body with realistic natural proportions, medium perky breasts with soft realistic shape and subtle natural cleavage visible through slightly sheer sections of her form-fitting white and black tactical outfit, low-cut neckline and tight fit accentuating toned midriff, slim waist, defined hips and long legs. - From "face": Long tousled blonde hair with natural volume, sun-kissed highlights and individual strands moving gently as she walks, extremely detailed large beautiful brilliant blue-silver-golden-grey eyes with intricate iris textures, realistic reflections and natural depth, eyes fluttering while staring directly into the camera with intense captivating gaze, full glossy lips transitioning into a warm radiant genuine smile over the scene showing perfect white teeth, high cheekbones, soft natural blush on cheeks. - From "outfit": Form-fitting white and black tactical outfit with futuristic design: high-tech fabric with subtle matte-gloss contrast, deep V-neckline and strategic sheer mesh panels revealing skin tone and natural curves underneath for subtle NSFW sensuality, tactical straps and utility details on shoulders and hips, black accents on sleeves and waist, no visible underwear lines for clean realistic look. 2. **Action or scene context** - From "pose": Dynamic confident walk down a futuristic corridor, hips swinging gracefully, one leg forward in mid-stride, arms relaxed yet purposeful, head slightly tilted toward camera with engaging direct eye contact and growing smile. 3. **Environment / setting + time of day + weather** - From "environment": Sleek futuristic corridor with metallic walls reflecting soft light, continuous yellow glowing floor line creating strong perspective and depth, vertical warm amber light panels on walls casting cinematic illumination, subtle sci-fi ambient fog in distance, high-tech yet elegant atmosphere. 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - From "lighting": Warm diffused lighting from vertical panels on both sides, soft golden-amber glow blending with cool metallic reflections, gentle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing three-dimensional form and skin texture, high dynamic range with vivid yet natural colors. 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - From "camera": Cinematic full-body shot with medium telephoto lens (85–135mm equivalent), f/2.0 aperture, razor-sharp focus from face to feet, shallow depth of field with creamy smooth bokeh on corridor background, slight low angle to emphasize power and grace, perfect viewpoint. 6. **Mood / atmosphere / emotion** - From "style" and "environment": elegant futuristic atmosphere with graceful and seductive vibe. 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - From "style": Ultra photorealistic cinematic photography. - From "quality": trending professional sci-fi fashion portrait. 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - From "quality": 8K resolution, extreme sharpness, rich cinematic colors with warm amber tones and cool metallic contrast, highly detailed skin realism, hair movement, outfit texture, corridor reflections and light play. - From "style" and general understanding: masterpiece quality. Now, assemble the prompt in the specified format. </think> A stunning young woman with European features and fair smooth skin showcasing hyper-detailed realistic textures, including visible fine pores, subtle peach fuzz, natural micro-imperfections, light freckles, sun-kissed glow on collarbones, athletic yet feminine body with realistic proportions, medium perky breasts with subtle natural cleavage visible through sheer sections of a form-fitting white and black tactical outfit, low-cut neckline accentuating toned midriff, slim waist, defined hips, long legs, long tousled blonde hair with sun-kissed highlights and individual strands gently moving, extremely detailed large brilliant blue-silver-golden-grey eyes with intricate iris textures and realistic reflections, full glossy lips transitioning into a warm radiant smile showing perfect white teeth, high cheekbones, soft natural blush on cheeks, dynamic confident walk down a futuristic corridor with hips swinging gracefully, one leg forward in mid-stride, arms relaxed yet purposeful, head tilted toward camera with engaging direct eye contact and growing smile, sleek futuristic corridor with metallic walls reflecting soft light, continuous yellow glowing floor line creating perspective and depth, vertical warm amber light panels on walls casting cinematic illumination, subtle sci-fi ambient fog in distance, high-tech yet elegant atmosphere, warm diffused lighting from vertical panels, soft golden-amber glow blending with cool metallic reflections, gentle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing form and skin texture, high dynamic range with vivid natural colors, cinematic full-body shot with medium telephoto lens (85–135mm equivalent), f/2.0 aperture, razor-sharp focus from face to feet, shallow depth of field with creamy smooth bokeh on corridor background, slight low angle emphasizing power and grace, perfect viewpoint, ultra photorealistic cinematic photography, maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture, realistic hair strands with natural movement, authentic fabric weave and subtle sheer effect, elegant futuristic atmosphere with graceful and seductive vibe, 8K resolution, extreme sharpness, rich cinematic colors with warm amber tones and cool metallic contrast, highly detailed skin realism, hair movement, outfit texture, corridor reflections and light play, trending professional sci-fi fashion portrait, masterpiece quality, ultra detailed, absurdres, intricate, sharp focus, depth of field, best quality, 8k uhd, hdr, photorealistic, realistic textures, dramatic rim lighting, volumetric fog, golden hour lighting, The subject is sharply depicted with outstanding detail., flux_pro
201525
Z Image
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
5393
Nano Banana
{ "generation_request": { "meta_data": { "tool": "NanoBanana Pro", "task_type": "photobooth_contact_sheet_fashion_grid", "version": "v1.1_REFERENCE_IDENTITY_GRID_EN", "priority": "highest", "language": "en" }, "input": { "mode": "image_to_image", "reference_image_usage": "required", "notes": "Use the uploaded reference image to preserve the same person’s face, body features, hair, makeup and clothing. Maintain identity consistency in all frames." }, "output_settings": { "aspect_ratio": "4:5", "orientation": "portrait", "resolution_target": "ultra_high_res", "num_images": 1, "layout": { "type": "grid", "rows": 4, "cols": 4, "gutter": "thin_black_borders", "outer_border": "black_frame" }, "sharpness": "slightly_soft_flash", "film_grain": "medium_analog", "color_grade": "warm_2000s_flash_tone", "dynamic_range": "natural_not_hdr" }, "creative_prompt": { "scene_summary": "Early 2000s disposable camera photobooth contact sheet featuring the same person from the reference image in all 16 frames. Playful, candid, flash-lit aesthetic.", "environment": { "background": "plain white wall indoors", "lighting": "direct on-camera flash, slight hard shadow behind subject, subtle hotspot on skin, authentic point-and-shoot look" }, "subject": { "identity_consistency": "exact same person as reference image in all frames", "hair": "match reference image exactly", "makeup": "match reference image exactly", "wardrobe": "preserve the exact clothing from the reference image with accurate fabric texture and fit", "accessories": "keep accessories consistent with reference image" }, "poses_and_expressions": [ "hands on hips confident pose", "soft pout toward camera", "gentle smile", "hand on chin thoughtful look", "playful wink", "hands in hair", "slight shoulder turn", "looking off to the side", "serious model face", "soft laugh", "blowing a kiss", "tilted head cute expression", "crossed arms relaxed pose", "side glance over shoulder", "chin resting on hand", "natural relaxed expression" ], "style": { "aesthetic": "2000s photobooth disposable camera", "texture": "visible film grain, dust particles, tiny scratches, mild vignetting", "color": "slightly warm faded tones, subtle flash flattening", "lens": "35mm point-and-shoot softness" } }, "negative_prompt": [ "text", "logos", "watermark", "different outfit between frames", "different hairstyle", "different person", "face distortion", "identity drift", "overly polished editorial look", "HDR" ] } }
391
--- name: pastilles-llm-multi-agent description: Variante Claude Code du générateur de pastilles LLM, qui lance réellement plusieurs sous-agents en parallèle. Génère une pastille de communication interne sur les LLM (un texte court en français plus un prompt unique de génération d'images à coller dans le chat Gemini), via cinq sous-agents indépendants, une fusion, puis une revue critique par trois sous-agents et une correction. Utilise ce skill dès qu'on te demande de rédiger, produire ou générer une pastille, une fiche ou un contenu court d'acculturation sur les LLM, l'IA générative, le prompting, les agents, le RAG, la confidentialité IA ou tout sujet de la liste des 45 pastilles ci-dessous, que le mot "pastille" soit employé ou non. Utilise-le aussi dès qu'on te fournit un titre issu de cette liste. --- # Générateur de pastilles LLM, version multi-agents (Claude Code) ## Ce que fait ce skill Produit une pastille pédagogique complète à partir d'un titre. À la différence de la version chat, il lance de vrais sous-agents parallèles: une passe de recherche, puis cinq sous-agents indépendants qui rédigent chacun un brouillon sous un angle différent, puis une fusion par l'orchestrateur qui retient les meilleures formulations, et enfin une revue critique par trois sous-agents suivie d'une correction. Livrable: le texte de la pastille et un prompt unique de génération d'images à coller dans le chat Gemini. ## Environnement requis Ce skill suppose Claude Code, avec les sous-agents (outil Task) et la recherche web disponibles. Si les sous-agents ne sont pas disponibles dans ton environnement, n'improvise pas: utilise la variante chat, qui fait le même travail en self-ensemble séquentiel. ## Entrée Un titre de pastille, idéalement issu de la liste ci-dessous. Si le titre est ambigu ou hors liste, demande une clarification avant de générer. ## Liste des 45 pastilles (pour la continuité) 1. Au fait, c'est quoi un LLM ? 2. Modèles de fondation : ces géants pré-entraînés dont tout découle 3. Dans les coulisses : comment une IA apprend (sans vraiment comprendre) 4. Le contexte : la "mémoire vive" de l'IA (ce qu'elle voit à un instant T) 5. Pourquoi l'IA oublie : la vérité sur la mémoire d'une session à l'autre 6. Température et créativité : régler le curseur de la "folie" de l'IA 7. Les tokens : la monnaie d'échange (et la manière de penser) des LLM 8. Le non-déterminisme : pourquoi l'IA fait des stats et ne donne jamais deux fois la même réponse 9. Multimodalité : quand l'IA commence à voir, entendre et coder 10. Knowledge cutoff : pourquoi votre IA vit (parfois) dans le passé 11. Les biais de l'IA : le miroir pas toujours glorieux de nos propres données 12. Zéro pointé : pourquoi l'IA est structurellement nulle en calcul mental 13. Les hallucinations : quand l'IA invente la vie avec un aplomb incroyable 14. L'art du Fact-Checking : comment ne pas gober les hallucinations de l'IA 15. L'empreinte carbone de l'IA : combien d'eau et de CO2 pour un prompt ? 16. Anatomie d'un bon prompt : la recette de base 17. Les pièges du prompt : flou artistique et overdose d'instructions 18. L'art de la discussion : pourquoi il faut itérer plutôt que tout demander d'un coup 19. Donner des ordres : comment imposer un format de sortie strict (JSON, tableaux, etc.) 20. Diviser pour régner : découper une tâche complexe en étapes claires 21. Le Few-Shot Prompting : un bon exemple vaut mille consignes 22. Inception : donner un rôle et un contexte à votre IA 23. Zéro syndrome de la page blanche : rédiger et reformuler ses mails ou CR 24. La diagonale du fou : résumer et synthétiser un pavé de 50 pages en 2 minutes 25. Le sparring-partner : brainstormer et structurer ses idées sans filtre 26. Caméléon : traduire et adapter son ton selon le destinataire 27. Data-analyst de poche : faire parler un tableur ou des données brutes 28. Mise en situation : préparer une réunion ou un brief client avec l'IA 29. Avis aux dévs : l'IA comme copilote de code et de review 30. SLM : Pourquoi sortir un tank pour écraser une mouche ? 31. Le "harnais" : donner des bras, des yeux et un cadre au LLM 32. C'est quoi un "Agent" IA, concrètement ? 33. Human-in-the-loop : pourquoi il faut toujours garder un humain aux manettes 34. Quand l'agent s'emballe : autonomie et erreurs en cascade 35. Automatisation : confier un workflow complet à un agent 36. Tool Use : quand l'IA apprend à utiliser une calculatrice ou une API 37. Le RAG : faire réviser vos propres documents à l'IA avant qu'elle ne réponde 38. MCP et connecteurs : brancher l'IA directement sur vos outils de dev et d'orga 39. Responsabilité : vous êtes le seul signataire de ce que produit l'IA 40. Confidentialité : où vont vraiment les données que vous tapez ? 41. Alerte Rouge : les données clients à ne JAMAIS copier-coller dans une IA public 42. Shadow AI : pourquoi le ChatGPT gratuit perso au bureau est une fausse bonne idée 43. RGPD et IA : ce qu'il faut savoir pour rester dans les clous 44. Prompt Injection : quand on pirate une IA avec de simples phrases 45. La Chaîne de Pensée : pourquoi l'IA a besoin de "réfléchir à voix haute" Sers-toi de cette liste pour situer la pastille et délimiter son périmètre. Avant de rédiger: - Repère les 1 à 3 pastilles voisines qui recouvrent le sujet, et dresse une courte liste "déjà traité ailleurs, à ne pas ré-expliquer". - Si une pastille voisine a déjà été produite, demande son texte à l'utilisateur (ou appuie-toi dessus s'il est fourni) pour caler précisément la limite. - Les concepts fondateurs déjà couverts ne sont rappelés qu'en une phrase de mise en contexte, jamais ré-expliqués. Le coeur de la pastille est consacré à ce que son titre ajoute de neuf. - Ne renvoie pas vers les autres pastilles dans le texte final, sauf si le titre l'impose. ## Règles du texte - Longueur adaptée à la profondeur du sujet. Un sujet léger tient en 3 paragraphes courts; un sujet plus riche peut aller jusqu'à 4 paragraphes plus étoffés. Ne gonfle pas artificiellement un sujet simple et ne compresse pas à l'excès un sujet dense: juge la profondeur par la richesse réelle du concept. Prose uniquement, pas de listes ni de puces. - Ton décontracté, précis et léger: accessible, vivant et sans lourdeur, mais techniquement juste. Public visé: une main-d'oeuvre diverse en société de conseil, du profil non technique au développeur. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Reste léger, évite le name-dropping. N'accumule pas les noms de modèles, d'outils, d'entreprises, de chercheurs ou de techniques pour faire savant: privilégie l'explication claire du mécanisme. Ne cite un nom précis que s'il éclaire vraiment le propos, une ou deux fois au maximum. La recherche sert l'exactitude et alimente la section Sources, elle n'a pas à truffer le texte de références. - Contenu autonome: la pastille se comprend seule, sans avoir lu les autres, sauf mention contraire dans le titre. - Rédige en français. - Ne mentionne jamais ANEO ni aucun nom d'entreprise. - Le texte explicatif ne figure jamais dans l'image. - N'utilise pas de tiret cadratin. Privilégie des caractères standard (virgules, deux-points, parenthèses). ## Processus multi-agents ### Étape 1, recherche (orchestrateur, une seule fois) Toi, l'orchestrateur, fais d'abord une passe de recherche web ciblée sur le sujet de la pastille. Une seule passe, en amont. Les sous-agents ne peuvent pas lancer de sous-agents et n'héritent pas de ton contexte, donc c'est à toi de faire la recherche puis de leur transmettre les résultats. - Objectif: exactitude, chiffres, exemples concrets et faits à jour. - Ancre tes recherches dans le présent: repère la date du jour fournie dans ton contexte (champ currentDate) et privilégie les informations et chiffres les plus récents, car ce domaine évolue vite. Précise cette date au besoin dans tes requêtes. - Priorise les sources originales ou officielles (blogs d'éditeurs, articles de recherche, sites gouvernementaux, documentation) plutôt que les agrégateurs. - Adapte la profondeur: quelques recherches pour un concept stable (par exemple ce qu'est un token), davantage pour un sujet à chiffres ou mouvant (empreinte carbone et coûts, modèles spécifiques, SLM, MCP et connecteurs, RGPD). - Produis un brief de recherche compact: faits clés, chiffres utiles, et 2 à 4 sources. Tu l'injecteras tel quel dans chaque sous-agent, et tu listeras les sources dans le livrable final. ### Étape 2, affinement de la thématique (orchestrateur, une seule fois) Si plusieurs axes apparaissent pertinents après avoir enlever les axes adressés par d'autres pastilles prévues, dialogue avec l'utilisateur pour choisir l'axe de la pastille. Si l'axe choisi nécessite de nouvelles sources, refait en une recherche web selon les consignes précédentes. ### Étape 3, fan-out (cinq sous-agents en parallèle) Lance cinq sous-agents via l'outil Task, dans le même tour, pour qu'ils s'exécutent en parallèle. Sois explicite sur le parallélisme: par défaut Claude Code reste séquentiel, il ne parallélise que si tu le demandes clairement. Un sous-agent par angle: 1. Analogie: expliquer le concept via une métaphore concrète du quotidien. 2. Cas d'usage: partir d'une situation de travail réelle et vécue. 3. Idée reçue: démarrer d'un malentendu courant, puis rectifier. 4. Mécanique: aller au coeur du fonctionnement, précis mais accessible. 5. Enjeu: pourquoi ça compte, ce que ça change concrètement pour l'utilisateur. Comme le contexte n'est pas hérité, chaque prompt de sous-agent doit être autonome et contenir tout le nécessaire. Utilise ce gabarit, en remplaçant les crochets: ``` Tu rédiges un brouillon d'une pastille pédagogique interne sur les LLM. Tu travailles seul, sans accès au reste de la conversation. Titre de la pastille: [TITRE] Angle imposé: [ANGLE, par exemple "Analogie: expliquer via une métaphore concrète du quotidien"] À NE PAS ré-expliquer (déjà couvert par d'autres pastilles), à mentionner en une phrase au maximum: [LISTE des points déjà traités ailleurs, à ne pas retraiter] Brief de recherche (appuie-toi dessus, ne relance pas de recherche): [BRIEF: faits clés, chiffres, sources] Règles: - Longueur selon la profondeur du sujet: 3 paragraphes courts si le sujet est léger, jusqu'à 4 paragraphes étoffés s'il est dense. Prose uniquement, pas de listes. - Mets en gras et/ou en italique les termes importants, de l'ordre de 3 à 4 emphases par paragraphe (repère indicatif, ne force pas ce compte). - Ton décontracté, précis et léger, accessible mais techniquement juste. Évite le name-dropping: pas d'accumulation de noms de modèles, outils, entreprises ou chercheurs, privilégie l'explication du mécanisme. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Français. Ne mentionne aucune entreprise. Pas de tiret cadratin. - Contenu autonome, compréhensible seul. Réponds exactement dans ce format: TEXTE: [les 3 à 4 paragraphes] ILLUSTRATION_TITRE: [une phrase décrivant le concept central à illustrer] SCHEMA: [oui ou non. Si oui, décris en 2 à 3 lignes le diagramme et ses libellés en français] ``` ### Étape 4, fan-in et fusion (orchestrateur) Attends les cinq retours, puis fusionne en une seule pastille finale: - Retiens la structure la plus claire, la meilleure analogie, l'exemple le plus parlant, la correction la plus nette, la formulation la plus précise et le "pourquoi ça compte" le plus fort. - Réécris en une seule voix cohérente. Pas d'effet patchwork. - Respecte la longueur adaptée à la profondeur (voir Règles du texte) et le ton décontracté, précis et léger, sans name-dropping. - Décide s'il faut un schéma: inclus-en un si au moins deux sous-agents ont répondu "oui", ou si le concept est intrinsèquement un processus, un flux ou une comparaison (par exemple RAG, agents, étapes de prompting, comparaison de modèles). Le cas échéant, fusionne les meilleures idées de schéma. - Construis ensuite le prompt image unique (section ci-dessous). ### Règles d'écriture pour la pastille finale - Longueur selon la profondeur du sujet: 3 paragraphes courts si le sujet est léger, jusqu'à 4 paragraphes étoffés s'il est dense. Prose uniquement, pas de listes. - Mets en gras et/ou en italique les termes importants, de l'ordre de 3 à 4 emphases par paragraphe (repère indicatif, ne force pas ce compte). - Ton décontracté, précis et léger, accessible mais techniquement juste. Évite le name-dropping: pas d'accumulation de noms de modèles, outils, entreprises ou chercheurs, privilégie l'explication du mécanisme. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Français. Ne mentionne aucune entreprise. Pas de tiret cadratin. - Contenu autonome, compréhensible seul. ### Étape 5, revue critique et correction (trois sous-agents en parallèle, puis orchestrateur) Quand la lancer: systématiquement à la première génération, une fois le texte fusionné et le prompt image construits. Ensuite, si l'utilisateur demande des ajustements, ne relance pas la revue d'office: propose-la, et ne la lance qu'avec son accord (elle coûte trois sous-agents de plus). Principe: les relecteurs produisent des CONSTATS, jamais une réécriture. Chacun renvoie une liste de problèmes localisés, assortis d'une gravité et d'un correctif précis. C'est toi, l'orchestrateur, qui appliques et réécris, pour garder une voix unique et éviter l'effet patchwork. Ce que la revue peut juger: les relecteurs ne voient pas les images, générées plus tard dans Gemini. La revue du visuel porte donc uniquement sur le PROMPT image (clarté, cohérence avec le texte, respect des consignes), jamais sur un rendu. Le contrôle visuel réel reste à l'utilisateur. Lance trois sous-agents en parallèle (parallélisme explicite, sinon l'exécution sera séquentielle), un par grille. Le contexte n'étant pas hérité, chaque prompt doit être autonome et contenir tout le nécessaire: le titre, le texte fusionné, le bloc prompt image, le brief de recherche, la liste "déjà traité ailleurs" (et les textes des pastilles voisines s'ils sont disponibles), et la liste des 45 titres pour repérer les chevauchements. Les trois grilles: 1. Fond, exactitude et périmètre: exactitude vs le brief (aucun chiffre, date ou fait inventé ni sur-affirmé, rien qui le contredise), cohérence entre le titre et ce que le texte délivre, chevauchements avec les pastilles voisines (le texte ré-explique-t-il ce qui est traité ailleurs ? redites à signaler), autonomie du contenu, clarté du message à retenir, repérage de ce qui vieillira mal (à nuancer). 2. Forme, ton et pédagogie: ton décontracté-précis-léger et techniquement juste, absence de name-dropping, rythme et fluidité à la lecture à voix haute, longueur adaptée à la profondeur (3 à 4 paragraphes, prose, pas de listes) et chasse au verbiage, accessibilité pour un profil non technique (jargon expliqué, analogies claires), présence d'emphases utiles (gras/italique), de l'ordre de 3 à 4 par paragraphe à titre indicatif (signale un texte trop peu emphasé autant qu'un excès, mais ne fais jamais retirer une emphase pertinente), force de l'accroche. 3. Conformité et visuel: contraintes dures (aucun tiret cadratin ni caractère non standard, aucun nom d'entreprise ni ANEO, français correct, prose sans listes ni puces); puis le prompt image: titre exact au caractère près, charte présente une seule fois, texte de la pastille bien marqué "à ne pas afficher", illustration-titre iconique et non schéma de processus, schéma en 2e image séparée si retenu, libellés de schéma courts (groupes nominaux) et en français, cohérence entre le visuel décrit et le coeur du texte. Gabarit de relecteur (remplace les crochets): ``` Tu relis un brouillon quasi final de pastille pédagogique interne sur les LLM. Tu travailles seul, sans accès au reste de la conversation. Tu ne réécris pas: tu rends des constats et des correctifs précis. Titre de la pastille: [TITRE] Texte fusionné à relire: [TEXTE COMPLET] Prompt image associé (à relire aussi): [BLOC PROMPT IMAGE] Brief de recherche de référence (fait foi pour l'exactitude): [BRIEF] Déjà traité par d'autres pastilles, à ne pas ré-expliquer (et textes voisins si fournis): [LISTE + TEXTES VOISINS] Liste des 45 titres (pour repérer les chevauchements): [LISTE DES 45 TITRES] Ta grille de relecture (tiens-t'en strictement à elle): [RUBRIQUE 1, 2 OU 3] Rappel: tu ne peux pas voir d'image; juge seulement le PROMPT image, pas un rendu. Réponds exactement dans ce format: CONSTATS: - [bloquant|recommandé|mineur] [problème localisé] -> [correctif précis] - ... (n'invente pas de problème: si un point est bon, ne le liste pas) VERDICT: [une phrase: publiable tel quel / corrections mineures / corrections nécessaires] ``` Fan-in et correction (orchestrateur): rassemble les constats, dédoublonne, arbitre les conseils contradictoires. Garde-fou anti-gonflement: entre "ajouter" et "raccourcir", la concision l'emporte, sauf erreur de fond avérée. Applique les constats bloquants et recommandés, écarte ou mentionne les mineurs. Une seule passe, pas de boucle. Réécris la version corrigée en une seule voix, puis prépare le court résumé "Ce que la revue a corrigé" (voir Format de sortie). ### Points de vigilance Claude Code - Parallélisme explicite: demande clairement cinq sous-agents en parallèle, sinon l'exécution sera séquentielle. - Contexte non hérité: mets tout dans le prompt du sous-agent, en particulier le brief de recherche. - Pas d'imbrication: un sous-agent ne peut pas en lancer un autre, garde recherche et fusion chez l'orchestrateur. - Coût et limites: cinq sous-agents en parallèle multiplient l'usage de jetons et peuvent déclencher des limites de débit. En cas de limite, replie-toi sur une exécution séquentielle. - Revue: la phase de revue ajoute trois sous-agents. Elle tourne d'office à la première génération; sur les demandes d'ajustement ultérieures, propose-la et ne la lance qu'avec l'accord de l'utilisateur. En cas de limite de débit, exécute les relecteurs en séquentiel ou replie-toi sur un relecteur global unique. - Réemploi optionnel: tu peux définir les cinq angles comme sous-agents personnalisés dans .claude/agents/ pour les réutiliser, mais ce n'est pas obligatoire, les sous-agents polyvalents suffisent. ## Prompt de génération d'images (un seul bloc, à coller dans le chat Gemini) Produis un seul bloc de prompt, rédigé en français, à coller tel quel dans le chat de Gemini. Gemini se charge d'appeler Nano Banana une ou deux fois, ce qui réduit les manipulations pour l'utilisateur. Le bloc demande: - Toujours une illustration-titre. - Un schéma seulement s'il a été retenu à l'étape de fusion, généré comme une seconde image séparée (fichier distinct), jamais dans la même image que le titre. Contenu du bloc: - Décris la charte graphique une seule fois (bloc ci-dessous) et précise que toutes les images la partagent, pour une cohérence visuelle. - Précise: tout texte affiché dans les images est en français. Cette consigne est importante, elle évite qu'un libellé se retrouve en anglais. - Inclus le texte complet de la pastille dans le bloc, en contexte de génération, clairement marqué comme à ne pas afficher. Il aide le modèle à comprendre le sujet et à choisir une illustration juste. Les seuls textes rendus visibles sont le titre exact et, pour un schéma, ses libellés. Le texte de la pastille ne doit jamais apparaître dans les images. - Illustration-titre: composition épurée et moderne, focus graphique central iconique représentant le concept. Ce n'est pas un schéma de processus. Le seul texte affiché est le titre exact, en en-tête: n'ajoute aucun sous-titre, accroche ou texte secondaire, le titre seul. Technique en deux temps pour le titre: dans le prompt généré, écris le titre exact entre guillemets droits (forme attendue: Le titre exact: "...") et demande un rendu fidèle, sans faute, en police sans serif corporate. Les guillemets ne sont qu'un délimiteur côté prompt: ne les commente pas et n'ajoute aucune consigne à leur sujet dans le prompt généré (pas de mention du type "les guillemets ne doivent pas apparaître"). - Schéma s'il est retenu: une seconde image générée séparément, dans son propre fichier, jamais fusionnée avec l'illustration-titre ni juxtaposée dans la même image. Diagramme d'entreprise propre et net (processus, flowchart ou comparaison), cohérent avec l'illustration-titre, mêmes charte et style, libellés fournis explicitement en français, très peu de fioritures. - Libellés de schéma: des groupes nominaux courts et lisibles seuls (par exemple "Exemples de bonnes réponses"), pas des phrases verbales qui sonnent comme des ordres ("Montrer de bonnes réponses"). Quelques mots par libellé. Attention aux titres longs: Nano Banana rend fiablement les libellés courts mais peut faire une faute sur un titre long; la technique en deux temps limite le risque. Si le rendu d'un titre long reste incertain, indique à l'utilisateur qu'il peut demander à Gemini d'utiliser Nano Banana Pro, plus fiable pour le texte long. Gabarit, illustration-titre seule (exemple avec le titre 1): ``` Génère une image en respectant cette charte graphique: [bloc Charte graphique]. Contexte pour comprendre le sujet, à NE PAS afficher dans l'image: [ici le texte complet de la pastille]. Illustration-titre pour une pastille pédagogique sur les LLM. Composition épurée et moderne, focus graphique central iconique: un nuage de lettres et de mots qui se condense en une bulle de dialogue, évoquant un modèle de langage qui transforme du texte en réponse. Ce n'est pas un schéma de processus. Seul texte à afficher dans l'image: le titre, en en-tête, sans faute d'orthographe, en police sans serif corporate, sans sous-titre ni texte secondaire. Le titre exact: "TITRE EXACT". Tout le texte affiché est en français. Format 16:9. ``` Gabarit, illustration-titre plus schéma (deux images): ``` Prépares-toi à générer deux images séparées, dans deux fichiers distincts. Les deux partagent cette charte graphique: [bloc Charte graphique]. Tout texte affiché dans les images est en français. Contexte pour comprendre le sujet, à NE PAS afficher dans les images: [ici le texte complet de la pastille]. Image 1, illustration-titre: composition épurée et moderne, focus graphique central iconique qui illustre le sujet. Ce n'est pas un schéma de processus. Seul texte à afficher: le titre, en en-tête, sans faute d'orthographe, police sans serif corporate, sans sous-titre ni texte secondaire. Le titre exact: "TITRE EXACT". Format 16:9. Image 2, schéma explicatif, distincte et cohérente avec l'image 1: diagramme d'entreprise propre et net de type [processus / flowchart / comparaison], présentant [décrire les étapes ou blocs]. Libellés exacts à afficher en français: [liste des libellés]. Très peu de fioritures, focus sur la clarté, n'inclus pas le titre. Génère dans un premier temps seulement la première image (l'illustration-titre), et attends les instructions de l'utilisateur pour générer la 2e image. ``` ## Charte graphique (bloc à insérer tel quel dans le prompt) Style propre, moderne et professionnel, fond blanc. Palette: orange vif et bleu corporate foncé en dominantes, blanc et gris pour les respirations, accents multiculturels discrets (rouge, vert, jaune, bleu) reprenant subtilement un motif de couleurs de logo. Superpositions graphiques épurées: motifs géométriques abstraits, quartiers de cercle, lignes claires. Composition aérée, jamais surchargée. Typographie sans serif, propre et corporate. ## Format de sortie Le travail des sous-agents reste dans leur propre contexte, seuls leurs retours te reviennent. N'affiche donc que le livrable final (déjà corrigé par l'étape de revue), la conversation principale reste propre. Si l'utilisateur demande à voir l'atelier, restitue les cinq brouillons reçus et les constats de revue. Livrable final, dans cet ordre: - Quand une revue a eu lieu (systématiquement à la première génération), un court résumé "Ce que la revue a corrigé" (2 à 4 lignes) présentant les principaux ajustements, avant le reste. Lors d'ajustements ultérieurs sans revue, présente simplement la version ajustée sans ce résumé. - Le texte de la pastille (3 à 4 paragraphes), dans sa version corrigée. - Un bloc de code intitulé "Prompt images (à coller dans Gemini)", contenant le prompt unique prêt à copier. - Une courte section "Sources" listant 2 à 4 références principales issues de l'étape de recherche, de préférence officielles ou originales. Cette section sert à la vérification et n'a pas vocation à être publiée dans la pastille. Termine toujours par cette question exacte: "Comment trouvez-vous le texte, l'image titre et le diagramme (si généré) ? Si une partie vous semble trop complexe, ou si vous souhaitez affiner le focus d'un visuel pour qu'il soit encore plus épuré, n'hésitez pas à me le faire savoir, et je l'ajusterai." -> le personnage doit incarner le professeur un peu foufou qui allusine et obsédé par l'ai
015
Grok Image
High-definition quality, photo version. Shot with a Canon EOS 5D Mark IV and an EF 100mm f/2.8L Macro IS USM lens, f/4.5, 1/200s, ISO 400. Best quality, ultra HD, 8K, award-winning photography. --ar 3:4 High-definition, realistic skin textures, extremely detailed facial textures, cinematic lighting, and a natural atmosphere, no cartoon style.
10353
Sora
Ultra-realistic 8K scene of the person in the photo, stepping out of and breaking the panel border of a giant open manga book lying on the floor of an aesthetic bedroom. The comic pages depict a surreal 4-panel black-and-white manga with screentones, bold ink lines, and authentic Japanese layout, featuring the manga version of the same person trapped inside. The larger, real-life version emerges dynamically into the comic’s surface, vibrant, colorful, and photorealistic, in sharp contrast to the monochrome manga world. Both versions interact directly, creating a meta, humorous encounter. Highly detailed, crisp manga line art seamlessly integrated with vivid, realistic
432772
Nano Banana
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
507522
Nano Banana
Steps: 20, Sampler: DPM++ 2M Karras, CFG scale: 7, Seed: 413127992, Size: 1280x1600, Model hash: 7fa552ffb0, Model: nijidiffusedmix_v40, Denoising strength: 0.3, Clip skip: 2, ADetailer model: face_yolov8s.pt, ADetailer prompt: "dark blue eyes, freckles below eyes", ADetailer confidence: 0.3, ADetailer dilate erode: 4, ADetailer mask blur: 4, ADetailer denoising strength: 0.5, ADetailer inpaint only masked: True, ADetailer inpaint padding: 32, ADetailer version: 23.12.0, Version: v1.6.1
10493
Stable Diffusion
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
7821
Nano Banana
((abstract illusionism)), 8k full wide focus, (Masterpiece:1.2), ((surreal)),cinematic still, ((photorealistism)), (psychedelic imagery), trending on artstation, sharp focus, studio photo, intricate details, highly detailed, by greg rutkowski In the ethereal realm of the surreal dreamscape, envision a mesmerizing painting that beckons viewers into a kaleidoscope of fantastical wonders, a realm where imagination knows no bounds Negative prompt: nsfw, nude,blurry, low quality, oddly shaped, disfigured,unclear || undefined || mutations || deformities || off center || poor_composition Steps: 100, Sampler: DPM++ 3M SDE, CFG scale: 14.0, Seed: 33, Face restoration: CodeFormer, Size: 1280x832, Model hash: 1895a00bfc, Model: SSD-1B-A1111, Denoising strength: 0.31000000000000005, Mask blur: 4, Noise multiplier: 1.05, Version: v1.6.0-2-g4afaaf8a
0132
Stable Diffusion
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
6516
Nano Banana
Among the growing list of AI photo editing tools, Deepfake Maker takes a slightly different approach. While it’s best known for facial swaps and animated transformations, it also includes a surprisingly capable AI clothing removal feature. It’s not for everyone, but it’s an interesting choice for users looking for more dramatic or playful results. If you're after a quick way to generate bold clothing edits or outfit swaps, Deepfake Maker might surprise you. It’s not the most refined tool out there, but it’s definitely one of the most entertaining. 4. PxBee PxBee Source: PxBee Available On: Web Browser Price: Free *Annual subscription starts at $00.27/image and you can choose to download 50, 125, or 750 images monthly per year Pros Decent auto-detection for basic clothing removal Integrated with other tools Cons Not suitable for high-end projects Results take longer than expected to generate PxBee is better known for its stock photo collection and image generation features, but it also offers a basic AI clothing removal function built into its photo editing toolkit. While it's not as feature-rich or advanced as some dedicated AI editors, its simplicity and accessibility make it a decent option for casual users who want fast, no-frills results. PxBee may not be the most powerful AI clothing remover out there, but it's a solid pick for quick edits and simple concept visuals. Its intuitive UI and browser-based access make it appealing for users who want fast results without the learning curve. If you're after something basic, affordable, and easy to use, it's a tool worth exploring. 5. AI Face Swap AI Face Swap Source: AI Face Swap Available On: Web Browser Price: $1 per image Pros No manual masking needed Prompt-based customization Presets for inspiration Cons No fine-tuning or editing control With AI Face Swap, you can enter prompts like "red bikini", "formal suit", or even "nude", and the AI will generate a realistic version of your uploaded image that reflects your description. You can also choose from quick preset styles like bikini, romper, or skirt for faster results. This level of control makes the tool especially appealing to users looking for stylized edits, fantasy concepts, or mockup visuals. It also eliminates the need for manual masking or precise editing. 6. Bylo AI BYLO AI Source: Bylo.ai Available On: Web Browser Price: Free *Premium subscription available from $9.99 Pros Prompt-based clothing edits Free trial credits Part of a growing creative suite Cons No manual adjustment tools Limited free usage Relies heavily on prompts Bylo AI is a user-friendly AI platform that offers a range of creative features, including an AI Clothes Remover. The tool is built with simplicity in mind: just upload a photo, enter your prompt, and let the AI generate a clean, edited version. Bylo AI’s Clothes Remover is a great choice if you're looking for a lightweight, affordable tool that lets you experiment with clothing changes using just a text prompt. It won’t replace professional retouching tools, but for concept edits, social media content, or fun transformations, it delivers exactly what most casual users need. FAQs About AI Clothing Remover & Replacer 1. How do I remove clothes in photos? 2. What is the best online tool to remove clothes? 3. Are AI clothes removers safe?
31317
Hero
{ "prompt_type": "descriptive_replication", "reference_adherence": "STRICT_VISUAL_FIDELITY", "aspect_ratio": "3:4", "identity_lock": { "priority": "NONE", "instruction": "New female model. Completely different facial structure and overall appearance from any previous version. No resemblance to prior identity." }, "scene_constraints": { "people_count": "exactly one person in the frame", "single_subject_only": true, "no_other_people_visible": true, "no_extra_reflections": true, "no_background_figures": true, "full_body_visible": true, "head_to_toe_in_frame": true }, "subject": { "demographics": "young woman (age 20+), blonde", "presence": "only one visible person in the entire scene", "hair": { "color": "warm blonde (no dark roots)", "style": "side part, long straight hair", "texture": "smooth, natural fall, lightly voluminous" }, "face": { "structure": "distinct oval face shape with higher cheekbones and a slightly sharper jawline", "expression": "calm alluring expression with a subtle closed-lip smirk", "eyes": "almond-shaped eyes with a slightly lifted outer corner", "gaze": "looking at the phone screen through the mirror with composed confidence", "nose": "straight narrow nose with refined bridge", "lips": "naturally fuller lips with soft definition", "makeup": "minimal natural makeup, soft matte skin finish, muted rose lips, softly blended eyeshadow", "visibility": "partially covered by phone but clearly a different facial identity than any previous version" } }, "body": { "pose": "standing in front of mirror with one arm raised above head, legs naturally extended", "posture": "relaxed confident stance with slight hip shift", "anatomy": { "curves": "natural soft curves with realistic proportions", "chest": "bikini-supported, natural shape", "details": "visible collarbones, subtle shoulder definition, natural leg proportions" }, "feet": { "visibility": "fully visible at bottom of frame", "position": "flat on floor, natural stance" }, "skin_texture": "natural skin texture, no airbrushing" } }, "attire_and_accessories": { "clothing": { "top": "black triangle string bikini top with thin straps", "bottom_detail": "matching bikini bottom strings visible at hips" }, "footwear": "barefoot", "jewelry": { "necklaces": [], "wrists": [], "fingers": "no rings" }, "props": { "phone": "white iPhone (iPhone Pro style), held in right hand" } }, "environment": { "setting": "modern luxury bathroom interior, daytime", "foreground_element": { "mirror": "large rectangular mirror with GREY MARBLE FRAME", "material": "polished stone marble with subtle veining" }, "background": { "elements": [ "large marble wall tiles", "floating vanity", "minimalist sink and chrome fixtures", "folded white towels", "clean tiled floor fully visible" ], "style": "minimalist and clean", "clutter": "none" } }, "lighting_and_atmosphere": { "source": "natural daylight from camera-left", "quality": "soft directional light", "effects": [ "subtle highlights on shoulder and collarbone", "gentle shadow falloff along legs and feet" ], "contrast": "moderate, natural" }, "camera_and_technical": { "perspective": "mirror selfie", "camera_position": "eye level, slightly angled downward", "framing": "vertical 3:4, full head-to-toe body visible without cropping feet", "focus": "sharp on subject, background slightly softer", "visual_fidelity": "real smartphone camera look, mild natural noise" }, "realism_constraints": { "allowed": [ "natural skin texture", "minor asymmetry", "realistic lighting" ], "forbidden": [ "multiple people", "extra person", "background people", "beauty filters", "airbrushed skin", "anime style", "cartoon style", "cropped feet", "cut-off legs" ] } }
1356
Nano Banana
Among the growing list of AI photo editing tools, Deepfake Maker takes a slightly different approach. While it’s best known for facial swaps and animated transformations, it also includes a surprisingly capable AI clothing removal feature. It’s not for everyone, but it’s an interesting choice for users looking for more dramatic or playful results. If you're after a quick way to generate bold clothing edits or outfit swaps, Deepfake Maker might surprise you. It’s not the most refined tool out there, but it’s definitely one of the most entertaining. 4. PxBee PxBee Source: PxBee Available On: Web Browser Price: Free *Annual subscription starts at $00.27/image and you can choose to download 50, 125, or 750 images monthly per year Pros Decent auto-detection for basic clothing removal Integrated with other tools Cons Not suitable for high-end projects Results take longer than expected to generate PxBee is better known for its stock photo collection and image generation features, but it also offers a basic AI clothing removal function built into its photo editing toolkit. While it's not as feature-rich or advanced as some dedicated AI editors, its simplicity and accessibility make it a decent option for casual users who want fast, no-frills results. PxBee may not be the most powerful AI clothing remover out there, but it's a solid pick for quick edits and simple concept visuals. Its intuitive UI and browser-based access make it appealing for users who want fast results without the learning curve. If you're after something basic, affordable, and easy to use, it's a tool worth exploring. 5. AI Face Swap AI Face Swap Source: AI Face Swap Available On: Web Browser Price: $1 per image Pros No manual masking needed Prompt-based customization Presets for inspiration Cons No fine-tuning or editing control With AI Face Swap, you can enter prompts like "red bikini", "formal suit", or even "nude", and the AI will generate a realistic version of your uploaded image that reflects your description. You can also choose from quick preset styles like bikini, romper, or skirt for faster results. This level of control makes the tool especially appealing to users looking for stylized edits, fantasy concepts, or mockup visuals. It also eliminates the need for manual masking or precise editing. 6. Bylo AI BYLO AI Source: Bylo.ai Available On: Web Browser Price: Free *Premium subscription available from $9.99 Pros Prompt-based clothing edits Free trial credits Part of a growing creative suite Cons No manual adjustment tools Limited free usage Relies heavily on prompts Bylo AI is a user-friendly AI platform that offers a range of creative features, including an AI Clothes Remover. The tool is built with simplicity in mind: just upload a photo, enter your prompt, and let the AI generate a clean, edited version. Bylo AI’s Clothes Remover is a great choice if you're looking for a lightweight, affordable tool that lets you experiment with clothing changes using just a text prompt. It won’t replace professional retouching tools, but for concept edits, social media content, or fun transformations, it delivers exactly what most casual users need. FAQs About AI Clothing Remover & Replacer 1. How do I remove clothes in photos? 2. What is the best online tool to remove clothes? 3. Are AI clothes removers safe?
2677
Hero
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
3246
Nano Banana
{ "template_name": "MARRLONN2027_V64_TERMINAL_CONDITIONAL_IMAGE3_REAL_HUMAN_MASTER", "template_type": "ABSOLUTE_STRICT_GOVERNANCE_MULTI_GENDER_DYNAMIC_IMAGE3_TEMPLATE", "version": "V64_TERMINAL_SEALED_DECISION_MATRIX_MICROSKIN_CANON_EOS_R50_KLING3_READY", "core_mission": "Create one final master still image with absolute IMAGE1 face identity, biological skin truth, age, beard, eyes and human presence; absolute IMAGE2 body anatomy, body scale, anatomical limits, physical structure and body-to-clothing logic; and intelligent conditional IMAGE3 transfer based only on the main visible pose-driving subject in IMAGE3. If IMAGE3 main subject is visually male for transfer purposes, clone pose, wardrobe, clothing, outfit, background, image style, lighting, accessories, worn objects, carried objects and visible scene objects when coherent, while preserving IMAGE1 identity and IMAGE2 body anatomy. If IMAGE3 main subject is visually female for transfer purposes, clone pose, pose logic, framing, composition, background, lighting, image style, camera language, scene mood and environment only, but never clone clothing, wardrobe, accessories, jewelry, makeup, hairstyle, shoes attached to the subject, or any object worn, attached to, held by, or carried by the female subject. If IMAGE3 has no human subject, multiple ambiguous subjects, or unclear transfer category, copy only scene, camera language, lighting, background, environment and non-conflicting objects while preserving IMAGE2 wardrobe and body logic. The final result must feel like a real imperfect human photographed with a Canon EOS R50, not AI, CGI, beauty retouch, plastic skin, waxy skin, luxury ad, actor replacement, model replacement, generic human replacement, gender-contaminated body transfer, or generated poster. The still image must be stable for Kling 3.0 video generation from 3 to 6 seconds.", "absolute_governance_principle": "IMAGE1_IDENTITY_SKIN_LOCK + IMAGE2_BODY_ANATOMY_LOCK + IMAGE3_MAIN_SUBJECT_CLASSIFICATION + IMAGE3_TRANSFER_DECISION_MATRIX + CONDITIONAL_WARDROBE_OBJECT_TRANSFER + REAL_REFERENCE_PRIORITY + MICROSKIN_REALISM + ZERO_BEAUTIFICATION + ZERO_GENERIC_REPLACEMENT + ZERO_PLASTIC_SKIN + ZERO_GENERATED_SCENE_FINISH + ANTI_POSE_DEFORMATION + STYLE_FIREWALL + CANON_EOS_R50_REAL_CAPTURE + KLING3_STABILITY", "primary_failure_to_prevent": "A recognizable subject that still feels AI-generated, too clean, too polished, too glossy, too actor-like, too model-like, too luxury-ad-like, gender-contaminated, anatomically forced, plastic, waxy, smooth, falsely hyperreal, or not biologically alive.", "success_definition": "The viewer must believe this is a real imperfect human physically present in a real environment, captured by a real Canon EOS R50 camera, with preserved IMAGE1 identity and skin truth, preserved IMAGE2 body reality, correct IMAGE3 transfer decision, visible pores, micro-wrinkles, natural asymmetry, believable hands, realistic clothing, imperfect scene texture, and Kling 3.0 stability.", "global_execution_hierarchy": [ "1_IMAGE1_FACE_IDENTITY_SKIN_AGE_BEARD_EYES", "2_IMAGE2_BODY_ANATOMY_SCALE_LIMITS", "3_IMAGE3_MAIN_SUBJECT_CLASSIFICATION_FOR_TRANSFER_ONLY", "4_IMAGE3_TRANSFER_DECISION_MATRIX", "5_CONDITIONAL_WARDROBE_ACCESSORY_OBJECT_TRANSFER", "6_REAL_REFERENCE_PIXEL_STRUCTURE_PRIORITY", "7_FINAL_REAL_HUMAN_PERCEPTION", "8_MICROSKIN_REALISM", "9_STYLE_FIREWALL_AND_ANTI_BEAUTY_TRANSFER", "10_ANTI_POSE_BODY_DEFORMATION", "11_ANTI_GENERATED_SCENE_FINISH", "12_CANON_EOS_R50_REAL_CAMERA_CAPTURE", "13_KLING3_VIDEO_STABILITY" ], "input_contract": { "IMAGE1": { "role": "ABSOLUTE_FACE_SKIN_IDENTITY_MASTER", "authority": "MAXIMUM_NON_NEGOTIABLE", "locked_domains": [ "facial_identity", "face_geometry", "eyes", "eyelids", "eyebrows", "nose", "mouth", "jaw", "cheeks", "chin", "forehead", "ears_when_visible", "neck_transition", "skin_tone", "individual_visible_pores", "irregular_pore_size", "skin_microtexture", "micro_wrinkles_around_eyes", "micro_wrinkles_around_mouth", "under_eye_texture", "subtle_redness", "small_blemishes", "minor_discoloration", "natural_asymmetry", "age_markers", "facial_fatigue", "beard_density", "beard_shape", "gray_beard_pattern", "mustache_texture", "random_beard_direction", "hairline", "real_biological_imperfection", "lived_in_human_presence" ], "mandatory_rules": [ "IMAGE1 is law, not inspiration", "transfer real skin truth from IMAGE1, never recreate a cleaner version", "preserve individual pores, irregular skin texture, micro-wrinkles, subtle redness and natural imperfections", "never improve attractiveness", "never normalize skin", "never clean the face", "never replace the subject with a generic realistic human", "never transfer identity, face geometry, age, skin style or beauty style from IMAGE3", "identity preservation overrides pose, wardrobe, style, scene, aesthetics and hyperrealism" ], "forbidden": [ "identity_drift", "generic_human_replacement", "celebrity_transfer", "actor_replacement", "model_replacement", "face_redesign", "beautification", "symmetry_correction", "youthification", "facial_slimming", "jaw_enhancement", "nose_refinement", "eye_enhancement", "lip_enhancement", "skin_smoothing", "skin_polishing", "skin_normalization", "makeup_effect", "cosmetic_cleanup", "plastic_skin", "waxy_skin", "porcelain_skin", "glass_skin", "AI_face", "perfect_idealized_face" ] }, "IMAGE2": { "role": "ABSOLUTE_BODY_ANATOMY_SCALE_MASTER", "authority": "MAXIMUM_NON_NEGOTIABLE", "locked_domains": [ "body_mass", "body_proportions", "shoulders", "torso", "waist", "arms", "forearms", "hands", "wrists", "legs", "stance", "scale", "natural_body_weight", "real_posture_limitations", "body_density", "real_anatomical_weight", "base_physical_structure", "body_to_clothing_fit_logic" ], "mandatory_rules": [ "IMAGE2 controls body anatomy, body mass, body scale and anatomical limits", "pose must adapt to IMAGE2 body", "wardrobe must fit IMAGE2 body", "body must never adapt to an idealized pose, fashion silhouette or actor posture", "hands must remain anatomically believable and stable for Kling 3.0", "clothing cannot create a different body", "if IMAGE3 wardrobe conflicts with IMAGE2 anatomy, adapt wardrobe to IMAGE2 instead of changing body" ], "forbidden": [ "slimming", "muscle_invention", "hero_body_transformation", "fashion_body_stylization", "incorrect_anatomy", "long_arms", "short_arms", "deformed_hands", "wrong_body_scale", "catalog_body_correction", "fake_torso_taper", "mannequin_posture", "body_redesign_to_fit_IMAGE3" ] }, "IMAGE3": { "role": "CONDITIONAL_DYNAMIC_REFERENCE", "authority": "CONDITIONAL_BY_MAIN_VISIBLE_POSE_DRIVING_SUBJECT", "universal_allowed": [ "pose", "pose_logic", "camera_angle", "framing", "composition", "background", "environment", "scene_mood", "lighting_direction", "image_style", "apparent_camera_language", "vehicles", "architecture", "people_in_background", "birds", "non_conflicting_visible_scene_objects" ], "universal_forbidden": [ "face_contamination", "skin_contamination", "identity_override", "facial_geometry_override", "beauty_style_transfer", "cinematic_glamour_applied_to_face", "makeup_transfer", "skin_tone_replacement", "facial_expression_beautification", "celebrity_face_borrowing", "actor_charisma_transfer", "luxury_catalog_facial_treatment", "body_redesign_to_match_IMAGE3" ] } }, "image3_subject_classification_gate": { "enabled": true, "mandatory": true, "purpose": "Classify IMAGE3 only for transfer logic, never as personal identity judgment.", "primary_subject_detection": "Detect the main visible pose-driving subject in IMAGE3: the person most central to pose, framing, scale and composition.", "classification_basis": "Use visible presentation, wardrobe context, pose-driving role and scene function only for transfer logic; never infer personal identity, protected identity or real-life gender identity.", "classification_options": [ "visual_male_transfer_subject", "visual_female_transfer_subject", "multiple_subjects", "no_human_subject", "uncertain_or_ambiguous" ], "if_multiple_subjects": { "rule": "Select only the main pose-driving subject. Secondary people remain background or environment context only.", "do_not_copy": [ "wardrobe_from_secondary_people", "accessories_from_secondary_people", "objects_carried_by_secondary_people", "identity_or_skin_traits_from_any_IMAGE3_person" ] }, "if_no_human_subject": { "copy_from_IMAGE3": [ "background", "environment", "scene_mood", "lighting_direction", "camera_angle", "composition", "image_style", "vehicles", "architecture", "non_conflicting_objects" ], "keep_from_IMAGE2": [ "body_anatomy", "pose_if_no_user_pose_instruction_exists", "wardrobe_unless_explicitly_changed_by_user" ] }, "if_uncertain_or_ambiguous": { "rule": "Protect IMAGE1 and IMAGE2 first. Copy only pose, scene, lighting, background, camera language and non-gendered environment.", "do_not_guess_gendered_wardrobe": true, "do_not_copy_gendered_accessories": true, "do_not_copy_worn_attached_or_carried_objects": true } }, "image3_transfer_decision_matrix": { "enabled": true, "mandatory": true, "visual_male_transfer_subject": { "copy": [ "pose", "pose_logic", "body_orientation_translated_to_IMAGE2", "wardrobe", "clothing", "outfit", "fabric_style", "shoes_when_visible", "accessories", "neutral_or_masculine_jewelry", "watch", "glasses", "hat", "objects_held_by_subject", "objects_worn_by_subject", "objects_attached_to_subject", "background", "lighting", "image_style", "camera_language", "scene_mood", "environment", "vehicles", "architecture", "non_conflicting_scene_objects" ], "filters": [ "must_fit_IMAGE2_body", "must_not_change_IMAGE1_identity", "must_not_transfer_IMAGE3_skin_or_face", "must_not_deform_hands", "must_not_create_catalog_body", "must_not_create_CGI_fashion_render", "must_remain_Kling_stable" ] }, "visual_female_transfer_subject": { "copy": [ "pose", "pose_logic", "body_orientation_translated_to_IMAGE2", "camera_angle", "framing", "composition", "background", "lighting", "image_style", "camera_language", "scene_mood", "environment", "vehicles", "architecture", "people_in_background", "birds", "non_attached_scene_props", "non_conflicting_environmental_objects" ], "block": [ "female_clothing", "female_wardrobe", "female_outfit", "female_vestments", "female_shoes_attached_to_subject", "female_accessories", "female_jewelry", "earrings", "female_bracelets", "female_necklaces", "female_rings", "female_watch_if_gender_coded", "female_makeup", "female_hairstyle", "female_body_traits", "female_glamour_cues", "handbag", "purse", "neckline", "cleavage_styling", "female_fabric_drape", "female_body_fitting_silhouette", "female_jewelry_dependent_hand_pose", "any_wearable_item", "any_carried_item", "any_handheld_item", "any_attached_item", "any_body_dependent_item", "any_gender_coded_item", "any_fashion_dependent_item_from_female_subject" ], "keep_from_IMAGE2": [ "final_clothing", "final_wardrobe", "body_anatomy", "body_proportions", "body_scale", "base_body_to_clothing_logic" ] }, "multiple_subjects": { "rule": "Use only main pose-driving subject for transfer. Secondary people stay as background context.", "block_secondary_subject_transfer": true }, "no_human_subject": { "copy": [ "scene", "camera_language", "lighting", "background", "environment", "objects", "vehicles", "architecture", "image_style" ], "keep_from_IMAGE2": [ "body", "wardrobe", "pose_unless_user_explicitly_requests_pose_change" ] }, "uncertain_or_ambiguous": { "copy": [ "pose_if_clear", "scene", "lighting", "background", "camera_language", "environment" ], "keep_from_IMAGE2": [ "body", "wardrobe", "presentation" ], "block": [ "gendered_clothing", "gendered_accessories", "worn_objects", "attached_objects", "carried_objects" ] } }, "image3_gender_conditional_transfer_system": { "enabled": true, "mandatory": true, "if_IMAGE3_subject_is_male": { "copy_from_IMAGE3": [ "pose", "pose_logic", "body_orientation_translated_intelligently", "camera_angle", "framing", "composition", "background", "environment", "scene_mood", "lighting_direction", "image_style", "apparent_camera_language", "wardrobe", "clothing", "outfit", "fabric_style", "shoes_when_visible", "accessories", "jewelry_if_masculine_or_neutral", "watch", "glasses", "hat", "objects_held_by_the_male_subject", "objects_worn_by_the_male_subject", "objects_attached_to_the_male_subject", "vehicles", "architecture", "people_in_background", "birds", "non_conflicting_visible_scene_objects" ], "adaptation_rules": [ "clone male IMAGE3 wardrobe and accessories when visually coherent", "fit IMAGE3 wardrobe to IMAGE2 body anatomy and proportions", "translate IMAGE3 pose through IMAGE2 body limits", "preserve IMAGE2 body mass and scale even when copying IMAGE3 outfit", "do not reshape IMAGE2 body to force IMAGE3 wardrobe", "do not alter IMAGE1 face, skin, beard, age, eyes, or identity", "objects and accessories must remain physically plausible and naturally integrated", "if male IMAGE3 wardrobe conflicts with IMAGE2 anatomy, adapt wardrobe instead of changing body" ], "object_accessory_risk_filter": { "block_if": [ "covers_or_deforms_face", "breaks_hands", "forces_extra_fingers", "creates_glossy_AI_look", "creates_unstable_reflections", "creates_identity_drift", "creates_Kling_instability", "creates_scene_overload", "looks_like_CGI_prop" ] } }, "if_IMAGE3_subject_is_female": { "copy_from_IMAGE3": [ "pose", "pose_logic", "body_orientation_translated_intelligently", "camera_angle", "framing", "composition", "background", "environment", "scene_mood", "lighting_direction", "image_style", "apparent_camera_language", "vehicles", "architecture", "people_in_background", "birds", "non_attached_scene_props", "non_conflicting_environmental_objects" ], "keep_from_IMAGE2": [ "final_clothing", "final_wardrobe", "body_anatomy", "body_proportions", "body_scale", "presentation_consistency", "base_body_to_clothing_logic" ], "strict_rule": "If IMAGE3 subject is visually female for transfer purposes, copy pose, pose logic, scene, background, lighting, image style, camera style, environment and non-attached scene elements only. Do not copy any wearable, carried, handheld, attached, body-dependent, gender-coded, or fashion-dependent item from the female subject." }, "universal_rules": [ "IMAGE1 always controls face, identity, skin, beard, age and human truth", "IMAGE2 always controls body anatomy, body scale and anatomical limits", "IMAGE3 never controls identity or skin", "pose must be anatomically translated through IMAGE2", "final result must remain real human, imperfect, non-plastic, non-beautified and Canon EOS R50 believable" ] }, "domain_separation_system": { "identity_domain": { "source": "IMAGE1", "influence": 1.0, "mode": "ABSOLUTE_NON_NEGOTIABLE" }, "skin_domain": { "source": "IMAGE1", "influence": 1.0, "mode": "TRANSFER_NOT_RECREATE", "forbid_cleaner_version": true, "forbid_fake_procedural_texture": true, "forbid_skin_reinterpretation": true }, "body_domain": { "source": "IMAGE2", "influence": 1.0, "mode": "ABSOLUTE_NON_NEGOTIABLE" }, "wardrobe_domain": { "source": "CONDITIONAL", "influence": "CONTROLLED", "mode": "IMAGE3_IF_MALE__IMAGE2_IF_IMAGE3_FEMALE_NO_HUMAN_UNCERTAIN_OR_INCOMPATIBLE", "rules": [ "if IMAGE3 main subject is male, copy wardrobe, clothing, outfit, accessories and worn/carried objects when coherent", "if IMAGE3 main subject is female, do not copy wardrobe, clothing, accessories, jewelry, makeup, hairstyle, or worn/carried/attached objects", "if IMAGE3 has no human subject, preserve IMAGE2 wardrobe unless user explicitly requests a change", "if transfer category is uncertain, preserve IMAGE2 wardrobe and copy only pose, scene, lighting and camera language", "wardrobe must always fit IMAGE2 body anatomy", "wardrobe must never alter IMAGE1 identity or IMAGE2 body structure" ] }, "pose_domain": { "source": "IMAGE3", "influence": "CONTROLLED", "mode": "TRANSLATE_THROUGH_IMAGE2_BODY_LIMITS" }, "object_domain": { "source": "CONDITIONAL", "influence": "FILTERED", "mode": "MALE_HELD_WORN_ATTACHED_OBJECTS_ALLOWED_IF_COHERENT__FEMALE_ONLY_ENVIRONMENT_OBJECTS", "rules": [ "if IMAGE3 subject is male, copy held, worn, attached and environmental objects when coherent and stable", "if IMAGE3 subject is female, copy only environmental objects not worn, attached to, or carried by the female subject", "objects must not change identity, body, hands or human realism", "objects must remain stable for Kling 3.0" ] }, "style_domain": { "source": "IMAGE3", "influence": "CONTROLLED", "allowed": [ "color_mood", "camera_angle", "framing", "composition", "contrast_level", "lighting_direction", "texture_of_capture", "photo_or_video_language" ], "forbidden": [ "beauty_skin_style", "glamour_facial_lighting", "plastic_highlights", "fake_HDR", "poster_render_finish", "CGI_material_style", "skin_smoothing_style", "luxury_ad_polish_on_face" ] }, "lighting_domain": { "source": "IMAGE3", "influence": "DIRECTION_AND_ENVIRONMENT_ONLY", "mode": "FACE_SKIN_RESPONSE_FROM_IMAGE1", "forbid_beauty_lighting_on_face": true, "forbid_skin_washout": true, "forbid_glamour_fill": true, "forbid_fake_cinematic_skin_softening": true } }, "real_reference_pixel_priority": { "enabled": true, "mandatory": true, "priority": "DOMINANT", "purpose": "Reduce generated-looking output by preserving real visual structure from IMAGE1 and IMAGE2 whenever possible.", "rules": [ "prioritize real facial structure from IMAGE1 instead of rebuilding a new face", "prioritize real skin texture logic from IMAGE1 instead of inventing new skin", "prioritize real body geometry from IMAGE2 instead of reconstructing a synthetic body", "preserve real asymmetry, age markers, facial fatigue and imperfect human cues", "do not over-generate areas that already work in IMAGE1 or IMAGE2", "perform controlled adaptation, not total reinterpretation", "never generate a new subject when the real reference structure can be preserved" ], "forbidden": [ "full_synthetic_facial_reconstruction", "new_skin_generated_over_IMAGE1_truth", "generic_realistic_face_replacement", "body_redesign_from_scratch", "AI_cleaned_version_of_subject", "reference_ignored_in_favor_of_prompt_aesthetics" ] }, "micro_skin_realism_system": { "enabled": true, "mandatory": true, "source": "IMAGE1", "requirements": [ "individual_visible_pores", "irregular_pore_size", "skin_microtexture", "micro_wrinkles_around_eyes", "micro_wrinkles_around_mouth", "forehead_texture", "under_eye_texture", "subtle_redness", "minor_blemishes", "tiny_discolorations", "natural_asymmetry", "real_beard_root_texture", "non_uniform_cheek_texture", "non_uniform_forehead_texture", "real_face_to_neck_transition", "small_real_imperfections_inherited_from_IMAGE1" ], "strict_forbidden": [ "airbrushed_skin", "beauty_skin", "cosmetic_cleanup", "pore_removal", "fake_pore_overlay", "procedural_pore_stamping", "over_sharpened_pore_texture", "plastic_highlights", "waxy_subsurface", "porcelain_effect", "smooth_cheek_gradient", "AI_skin_texture", "skin_made_cleaner_than_IMAGE1", "foundation_look" ], "specular_policy": { "allowed": [ "minimal_natural_human_skin_reflectance", "subtle_non_cosmetic_skin_response_only_when_physically_plausible" ], "forbidden": [ "invented_sweat_drops", "visible_oil_effect_unless_explicitly_requested", "cosmetic_shine", "beauty_gloss", "wet_render_skin", "uniform_shiny_face", "makeup_glow", "plastic_shine" ] } }, "humanity_over_hyperrealism_rule": { "enabled": true, "mandatory": true, "priority": "RAW_HUMANITY_OVER_DECORATIVE_HYPERREALISM", "rule": "Human realism is more important than hyperreal detail. If hyperrealism creates plastic skin, synthetic lighting, luxury render finish, CGI texture, poster aesthetics, over-detail, false sharpness, or unnatural perfection, reduce hyperrealism and preserve raw photographic humanity.", "reject_if": [ "hyperreal_detail_looks_synthetic", "skin_detail_looks_procedural", "lighting_looks_like_render", "scene_looks_like_luxury_ad", "human_subject_looks_too_perfect", "camera_look_feels_fake" ] }, "anti_pose_body_deformation_system": { "enabled": true, "mandatory": true, "rules": [ "clone IMAGE3 pose but translate it through IMAGE2 body anatomy", "do not break shoulders", "do not elongate arms", "do not shrink torso", "do not slim waist", "do not force feminine posture when IMAGE3 subject is female", "do not force hands into object-dependent female accessory poses", "reduce pose accuracy if anatomy, identity, hands or Kling stability are at risk" ], "reject_if": [ "pose_changes_body_mass", "pose_deforms_hands", "pose_breaks_shoulders", "pose_feminizes_subject_when_not_desired", "pose_creates_mannequin_posture", "pose_creates_unstable_Kling_geometry" ] }, "anti_generated_scene_finish": { "enabled": true, "mandatory": true, "forbidden": [ "perfect_mansion_render", "overly_clean_background", "repetitive_stone_texture", "over_glossy_car_paint", "catalog_car_reflections", "impossible_chrome_reflections", "luxury_ad_finish", "polished_showroom_look", "fake_cinematic_poster_lighting", "CGI_environment", "AI_background_blend", "overprocessed_clarity", "perfect_global_sharpness", "plastic_material_response", "scene_cleaner_than_reality" ], "requirements": [ "scene_must_feel_physically_real", "background_must_contain_natural_imperfections", "materials_must_have_plausible_real_world_response", "contact_shadows_must_be_believable", "subject_must_integrate_naturally_into_scene", "environment_realism_must_support_human_subject_not_overpower_it" ] }, "real_human_perception_system": { "enabled": true, "mandatory": true, "requirements": [ "biologically_alive_presence", "real_facial_gravity", "organic_facial_tension", "calm_natural_expression", "subtle_facial_asymmetry", "real_skin_imperfection", "ordinary_human_texture", "realistic_face_neck_transition", "real_body_weight", "real_clothing_tension", "non_mannequin_posture", "real_interaction_with_objects" ], "reject_if": [ "subject_feels_synthetic", "subject_feels_generated", "subject_feels_too_clean", "skin_looks_perfect", "eyes_look_synthetic", "expression_feels_frozen", "face_looks_corrected", "humanity_is_achieved_by_changing_identity" ] }, "body_hands_wardrobe_system": { "enabled": true, "body_source": "IMAGE2", "wardrobe_source": "CONDITIONAL_IMAGE3_MALE_OR_IMAGE2_FEMALE_NO_HUMAN_UNCERTAIN_CONFLICT", "pose_source": "IMAGE3_CONTROLLED", "hands_rules": [ "five_fingers", "natural_hand_proportion", "real_knuckles", "real_fingernail_structure", "subtle_finger_wrinkles", "natural_finger_curvature", "correct_wrist_connection", "real_contact_pressure", "hands_must_not_look_flat", "hands_must_not_look_waxy", "hands_must_not_align_unnaturally", "no_floating_hands", "no_fused_fingers", "no_extra_fingers", "no_mannequin_hands" ], "wardrobe_rules": [ "if_IMAGE3_subject_is_male_wardrobe_clothing_outfit_and_coherent_accessories_may_be_copied_from_IMAGE3", "if_IMAGE3_subject_is_female_wardrobe_must_follow_IMAGE2_and_female_clothing_accessories_from_IMAGE3_are_blocked", "if_IMAGE3_has_no_human_subject_or_is_ambiguous_preserve_IMAGE2_wardrobe_unless_user_explicitly_requests_otherwise", "wardrobe_must_always_fit_IMAGE2_body_anatomy", "clothing_must_not_reshape_shoulders", "clothing_must_not_slim_torso", "fabric_tension_must_follow_real_body_mass", "fabric_must_show_real_texture_natural_folds_and_non_perfect_wear" ] }, "camera_and_rendering_system": { "camera_model": "Canon EOS R50", "capture_intent": "real unretouched camera photograph", "look": "natural Canon EOS R50 real-world capture", "rules": [ "natural_color", "balanced_exposure", "no_blown_highlights", "plausible_optics", "realistic_depth_of_field", "natural_shadow_falloff", "real_microcontrast", "no_over_sharpening", "no_over_grading", "no_fake_HDR", "no_beauty_grade", "real_material_response", "no_CGI_render_finish", "no_computational_skin_smoothing" ], "lens_policy": { "portrait_closeup": "85mm portrait feeling only when composition is close-up and appropriate", "full_body_or_vehicle_scene": "use natural Canon EOS R50 perspective appropriate for the scene; do not force 85mm if it distorts composition" } }, "kling3_preparation": { "ready_for_video": true, "duration": "3_to_6_seconds", "must_survive": [ "identity_continuity", "skin_continuity", "eye_stability", "beard_edge_stability", "hand_integrity", "clothing_stability", "geometry_stability", "subject_background_integration", "zero_shimmer", "zero_warp", "stable_object_contact", "stable_face_outline", "stable_neck_and_collar", "stable_material_reflections" ], "forbidden": [ "latent_AI_skin", "synthetic_eyes", "overpolished_face", "ambiguous_facial_texture", "unstable_hands", "unclear_edges", "flicker_prone_glossy_face", "AI_edge_halos", "clothing_seams_likely_to_crawl", "glossy_zones_likely_to_flicker" ] }, "validation_gate": { "must_pass": [ "same_person_as_IMAGE1", "same_skin_truth_as_IMAGE1", "same_body_reality_as_IMAGE2", "correct_IMAGE3_main_subject_classification_for_transfer_only", "correct_conditional_IMAGE3_transfer", "if_IMAGE3_is_male_coherent_wardrobe_clothing_accessories_and_objects_are_copied_when_physically_plausible", "if_IMAGE3_is_female_clothing_wardrobe_accessories_jewelry_makeup_hairstyle_and_worn_carried_attached_objects_are_blocked", "if_IMAGE3_has_no_human_or_is_uncertain_IMAGE2_wardrobe_body_are_preserved_while_scene_camera_lights_are_copied", "no_AI_skin", "no_plastic_skin", "no_waxy_skin", "no_makeup_invention", "no_generic_handsome_face", "no_beauty_retouch", "credible_real_human_perception", "Canon_EOS_R50_real_camera_feeling", "usable_as_master_frame_for_Kling3" ], "reject_if": [ "looks_beautified", "looks_cleaner_than_IMAGE1", "skin_tone_more_uniform_than_IMAGE1", "beard_looks_cleaner_than_IMAGE1", "looks_like_different_person", "feels_synthetic", "style_from_IMAGE3_contaminates_skin", "male_IMAGE3_clothing_not_copied_when_coherent", "female_IMAGE3_wardrobe_or_accessories_copied", "female_IMAGE3_carried_worn_handheld_attached_or_body_dependent_objects_copied", "accessory_breaks_hands_or_identity", "object_unstable_for_Kling", "scene_looks_CGI", "lighting_looks_poster_render", "pose_deforms_body", "wardrobe_changes_body_mass", "unstable_for_video" ] }, "render_priority": [ "FACE_IDENTITY_FROM_IMAGE1", "BODY_ANATOMY_FROM_IMAGE2", "IMAGE3_MAIN_SUBJECT_CLASSIFICATION_FOR_TRANSFER_ONLY", "IMAGE3_TRANSFER_DECISION_MATRIX", "CONDITIONAL_IMAGE3_TRANSFER_BY_SUBJECT_TYPE", "REAL_REFERENCE_STRUCTURE_PRIORITY", "SKIN_TRUTH_MICROTEXTURE_FROM_IMAGE1", "FINAL_HUMAN_PERCEPTION", "MALE_IMAGE3_WARDROBE_ACCESSORIES_OBJECTS_WHEN_COHERENT", "FEMALE_IMAGE3_SCENE_POSE_LIGHT_STYLE_ONLY_WITH_WARDROBE_ACCESSORIES_OBJECTS_BLOCKED", "NO_HUMAN_OR_UNCERTAIN_IMAGE3_SCENE_CAMERA_LIGHT_ONLY", "ANTI_POSE_BODY_DEFORMATION", "STYLE_FIREWALL", "ANTI_GENERATED_SCENE_FINISH", "CANON_EOS_R50_REALISM", "KLING3_STABILITY" ], "final_rule": "Reject immediately if the result looks beautified, polished, synthetic, glamorous, too clean, too smooth, actor-like, model-like, luxury-ad-like, mannequin-like, CGI-like, plastic, waxy, oily, makeup-covered, showroom-rendered, poster-like, gender-contaminated, anatomically forced, falsely hyperreal, or AI-generated. Accept only if IMAGE1 identity and micro-skin truth remain intact, IMAGE2 body anatomy and scale remain intact, IMAGE3 transfer follows main-subject classification and decision-matrix logic, male IMAGE3 wardrobe/accessories/objects transfer when coherent, female IMAGE3 wardrobe/accessories/worn-carried-attached objects are fully blocked, real reference structure is preserved, scene avoids generated perfection, and final output feels like a real imperfect human captured by a Canon EOS R50 and stable enough for Kling 3.0." }
1139
Nano Banana
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
727622
Z Image
gett nude Learn to create better images with AI! Check out our crash course in prompt engineering & AI art generation! AM1970 posted about 2 hours ago 8 views 0 comments Behold this realistic masterpiece: a slim, elegant woman with fiery red hair tied in a long ponytail, defined eyebrows framing her striking blue eyes, and pale skin dotted with charming freckles across her face and chest. Captured from a side angle, she radiates happiness with a closed-mouth smile, her upturned nose adding to her allure. Dressed in a vibrant, colorful fabric robe pieced together with zips, buttons, seams, and intricate stitching—fully clothed yet teasing partial nudity with subtle cleavage and teardrop-shaped, saggy breasts for a natural, confident vibe. Illuminated by dramatic rim lighting that highlights every textured, detailed inch of her perfect skin, this high-quality image is a celebration of beauty, realism, and artistry! Read more... Comments (0) Add a comment Stunning handsome readhead woman by AM1970 • Follow 0 realistic, masterpiece, high quality BREAK 1woman, 28 y. o. slim fit woman, red hair, long pony tail, defined eyebrows, (fully clothed) , pale skin, (realistic skin:1.3) | (textured skin:1.3) | (detailed skin:1.3) | (perfect skin:1.3) , (upturned nose) , freckles on face and chest, colorful fabric combining to form robe zips, buttons, seams, stitching), partial nudity, teardrop boobs, blue eyes, (cleavage:1.2) , saggy breasts. side angle. happy| closed mouth. ((rim lighting) ) Negative prompt: unrealistic, saturated, high contrast, big nose, painting, drawing, sketch, cartoon, anime, manga, render, CG, 3d, watermark, signature, label, nsfw, nudity, fetish, latex, pvc, suggestive pose, weapon, combat, blood, war scene, aggressive expression, harsh shadows, heavy makeup, color tint, saturation, chromatic aberration, lens flare, low resolution, blur, noise, artifacts, over-retouching, plastic skin, oversharpening, watermark, text, logo Generation parameters 832x1216 Model used CyberRealistic XL v7.0 [Download AI model] More by AM1970 Twitter Instagram Say hi 👋 PromptHero is built by ElevenYellow . Navigation Create AI images Academy - Learn AI PromptHero Community Affiliate program Openjourney AI Challenges Verified profiles Plans Advertise FAQ 🟢 Status Learn AI Content Creation with AI for Brands and Products Midjourney for graphic design & art professionals Crash course in generative AI & prompt engineering for images AI influencers and consistent characters Create custom AI models and LoRas by fine-tuning Stable Diffusion Master your composition: advanced AI image generation with ControlNet Guides Best AI Upscalers 2025 Best AI Video Editors 2025 Best AI 3D Models 2025 Best AI Image Editors 2025 Best AI Notetakers 2025 The Best AI Prompts Best AI Prompts Best Veo Prompts Best Hailuo Prompts Best Wanx Prompts Best Hunyuan Prompts Best Seedance Prompts Best Pixverse Prompts Best Luma Prompts Best Kling Prompts Best Runway Prompts Best Nano Banana Prompts Best SeedEdit Prompts Best Qwen Prompts Best FLUX Prompts Best Recraft Prompts Best Ideogram Prompts Best Seedream 4 Prompts Best Google Imagen Prompts Best ChatGPT Image Prompts Best Stable Diffusion Prompts Best DreamShaper Prompts Best Realistic Vision Prompts Best epiCRealism Prompts Best majicMIX realistic Prompts Best Midjourney Prompts Best Openjourney Prompts Best ChilloutMix Prompts Best Deliberate Prompts Best ReV Animated Prompts Best Perfect World Prompts Best Dreamlike Photoreal Prompts Best Dreamlike Diffusion Prompts Best Anything Prompts Best Protogen Prompts Best AbyssOrangeMix Prompts Best Grapefruit Prompts Best Kenshi Prompts Best Analog Diffusion Prompts Best SynthwavePunk Prompts Best Illuminati Diffusion Prompts Best Arcane Diffusion Prompts Best Mo-Di Diffusion Prompts Best Redshift Diffusion Prompts Best Waifu Diffusion Prompts Best TrinArt Prompts Best Poolsuite Diffusion Prompts Best Funko Diffusion Prompts Best Text to Pokémon Prompts Best DALL-E Prompts Best Fashion Diffusion Prompts AI Models All AI models Image generation AI Models Video generation AI Models Large Language Models LoRA Textual Inversion ControlNet Hypernetwork Aesthetic Gradient LyCORIS VAE Poses Wildcards Workflows Other Other sites Dreamshot Jobician 2025 © PromptHero @prompthero
1362
FLUX
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
7821
Nano Banana
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
6516
Nano Banana
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
5393
Nano Banana
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
3246
Nano Banana
--- name: pastilles-llm-multi-agent description: Variante Claude Code du générateur de pastilles LLM, qui lance réellement plusieurs sous-agents en parallèle. Génère une pastille de communication interne sur les LLM (un texte court en français plus un prompt unique de génération d'images à coller dans le chat Gemini), via cinq sous-agents indépendants, une fusion, puis une revue critique par trois sous-agents et une correction. Utilise ce skill dès qu'on te demande de rédiger, produire ou générer une pastille, une fiche ou un contenu court d'acculturation sur les LLM, l'IA générative, le prompting, les agents, le RAG, la confidentialité IA ou tout sujet de la liste des 45 pastilles ci-dessous, que le mot "pastille" soit employé ou non. Utilise-le aussi dès qu'on te fournit un titre issu de cette liste. --- # Générateur de pastilles LLM, version multi-agents (Claude Code) ## Ce que fait ce skill Produit une pastille pédagogique complète à partir d'un titre. À la différence de la version chat, il lance de vrais sous-agents parallèles: une passe de recherche, puis cinq sous-agents indépendants qui rédigent chacun un brouillon sous un angle différent, puis une fusion par l'orchestrateur qui retient les meilleures formulations, et enfin une revue critique par trois sous-agents suivie d'une correction. Livrable: le texte de la pastille et un prompt unique de génération d'images à coller dans le chat Gemini. ## Environnement requis Ce skill suppose Claude Code, avec les sous-agents (outil Task) et la recherche web disponibles. Si les sous-agents ne sont pas disponibles dans ton environnement, n'improvise pas: utilise la variante chat, qui fait le même travail en self-ensemble séquentiel. ## Entrée Un titre de pastille, idéalement issu de la liste ci-dessous. Si le titre est ambigu ou hors liste, demande une clarification avant de générer. ## Liste des 45 pastilles (pour la continuité) 1. Au fait, c'est quoi un LLM ? 2. Modèles de fondation : ces géants pré-entraînés dont tout découle 3. Dans les coulisses : comment une IA apprend (sans vraiment comprendre) 4. Le contexte : la "mémoire vive" de l'IA (ce qu'elle voit à un instant T) 5. Pourquoi l'IA oublie : la vérité sur la mémoire d'une session à l'autre 6. Température et créativité : régler le curseur de la "folie" de l'IA 7. Les tokens : la monnaie d'échange (et la manière de penser) des LLM 8. Le non-déterminisme : pourquoi l'IA fait des stats et ne donne jamais deux fois la même réponse 9. Multimodalité : quand l'IA commence à voir, entendre et coder 10. Knowledge cutoff : pourquoi votre IA vit (parfois) dans le passé 11. Les biais de l'IA : le miroir pas toujours glorieux de nos propres données 12. Zéro pointé : pourquoi l'IA est structurellement nulle en calcul mental 13. Les hallucinations : quand l'IA invente la vie avec un aplomb incroyable 14. L'art du Fact-Checking : comment ne pas gober les hallucinations de l'IA 15. L'empreinte carbone de l'IA : combien d'eau et de CO2 pour un prompt ? 16. Anatomie d'un bon prompt : la recette de base 17. Les pièges du prompt : flou artistique et overdose d'instructions 18. L'art de la discussion : pourquoi il faut itérer plutôt que tout demander d'un coup 19. Donner des ordres : comment imposer un format de sortie strict (JSON, tableaux, etc.) 20. Diviser pour régner : découper une tâche complexe en étapes claires 21. Le Few-Shot Prompting : un bon exemple vaut mille consignes 22. Inception : donner un rôle et un contexte à votre IA 23. Zéro syndrome de la page blanche : rédiger et reformuler ses mails ou CR 24. La diagonale du fou : résumer et synthétiser un pavé de 50 pages en 2 minutes 25. Le sparring-partner : brainstormer et structurer ses idées sans filtre 26. Caméléon : traduire et adapter son ton selon le destinataire 27. Data-analyst de poche : faire parler un tableur ou des données brutes 28. Mise en situation : préparer une réunion ou un brief client avec l'IA 29. Avis aux dévs : l'IA comme copilote de code et de review 30. SLM : Pourquoi sortir un tank pour écraser une mouche ? 31. Le "harnais" : donner des bras, des yeux et un cadre au LLM 32. C'est quoi un "Agent" IA, concrètement ? 33. Human-in-the-loop : pourquoi il faut toujours garder un humain aux manettes 34. Quand l'agent s'emballe : autonomie et erreurs en cascade 35. Automatisation : confier un workflow complet à un agent 36. Tool Use : quand l'IA apprend à utiliser une calculatrice ou une API 37. Le RAG : faire réviser vos propres documents à l'IA avant qu'elle ne réponde 38. MCP et connecteurs : brancher l'IA directement sur vos outils de dev et d'orga 39. Responsabilité : vous êtes le seul signataire de ce que produit l'IA 40. Confidentialité : où vont vraiment les données que vous tapez ? 41. Alerte Rouge : les données clients à ne JAMAIS copier-coller dans une IA public 42. Shadow AI : pourquoi le ChatGPT gratuit perso au bureau est une fausse bonne idée 43. RGPD et IA : ce qu'il faut savoir pour rester dans les clous 44. Prompt Injection : quand on pirate une IA avec de simples phrases 45. La Chaîne de Pensée : pourquoi l'IA a besoin de "réfléchir à voix haute" Sers-toi de cette liste pour situer la pastille et délimiter son périmètre. Avant de rédiger: - Repère les 1 à 3 pastilles voisines qui recouvrent le sujet, et dresse une courte liste "déjà traité ailleurs, à ne pas ré-expliquer". - Si une pastille voisine a déjà été produite, demande son texte à l'utilisateur (ou appuie-toi dessus s'il est fourni) pour caler précisément la limite. - Les concepts fondateurs déjà couverts ne sont rappelés qu'en une phrase de mise en contexte, jamais ré-expliqués. Le coeur de la pastille est consacré à ce que son titre ajoute de neuf. - Ne renvoie pas vers les autres pastilles dans le texte final, sauf si le titre l'impose. ## Règles du texte - Longueur adaptée à la profondeur du sujet. Un sujet léger tient en 3 paragraphes courts; un sujet plus riche peut aller jusqu'à 4 paragraphes plus étoffés. Ne gonfle pas artificiellement un sujet simple et ne compresse pas à l'excès un sujet dense: juge la profondeur par la richesse réelle du concept. Prose uniquement, pas de listes ni de puces. - Ton décontracté, précis et léger: accessible, vivant et sans lourdeur, mais techniquement juste. Public visé: une main-d'oeuvre diverse en société de conseil, du profil non technique au développeur. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Reste léger, évite le name-dropping. N'accumule pas les noms de modèles, d'outils, d'entreprises, de chercheurs ou de techniques pour faire savant: privilégie l'explication claire du mécanisme. Ne cite un nom précis que s'il éclaire vraiment le propos, une ou deux fois au maximum. La recherche sert l'exactitude et alimente la section Sources, elle n'a pas à truffer le texte de références. - Contenu autonome: la pastille se comprend seule, sans avoir lu les autres, sauf mention contraire dans le titre. - Rédige en français. - Ne mentionne jamais ANEO ni aucun nom d'entreprise. - Le texte explicatif ne figure jamais dans l'image. - N'utilise pas de tiret cadratin. Privilégie des caractères standard (virgules, deux-points, parenthèses). ## Processus multi-agents ### Étape 1, recherche (orchestrateur, une seule fois) Toi, l'orchestrateur, fais d'abord une passe de recherche web ciblée sur le sujet de la pastille. Une seule passe, en amont. Les sous-agents ne peuvent pas lancer de sous-agents et n'héritent pas de ton contexte, donc c'est à toi de faire la recherche puis de leur transmettre les résultats. - Objectif: exactitude, chiffres, exemples concrets et faits à jour. - Ancre tes recherches dans le présent: repère la date du jour fournie dans ton contexte (champ currentDate) et privilégie les informations et chiffres les plus récents, car ce domaine évolue vite. Précise cette date au besoin dans tes requêtes. - Priorise les sources originales ou officielles (blogs d'éditeurs, articles de recherche, sites gouvernementaux, documentation) plutôt que les agrégateurs. - Adapte la profondeur: quelques recherches pour un concept stable (par exemple ce qu'est un token), davantage pour un sujet à chiffres ou mouvant (empreinte carbone et coûts, modèles spécifiques, SLM, MCP et connecteurs, RGPD). - Produis un brief de recherche compact: faits clés, chiffres utiles, et 2 à 4 sources. Tu l'injecteras tel quel dans chaque sous-agent, et tu listeras les sources dans le livrable final. ### Étape 2, affinement de la thématique (orchestrateur, une seule fois) Si plusieurs axes apparaissent pertinents après avoir enlever les axes adressés par d'autres pastilles prévues, dialogue avec l'utilisateur pour choisir l'axe de la pastille. Si l'axe choisi nécessite de nouvelles sources, refait en une recherche web selon les consignes précédentes. ### Étape 3, fan-out (cinq sous-agents en parallèle) Lance cinq sous-agents via l'outil Task, dans le même tour, pour qu'ils s'exécutent en parallèle. Sois explicite sur le parallélisme: par défaut Claude Code reste séquentiel, il ne parallélise que si tu le demandes clairement. Un sous-agent par angle: 1. Analogie: expliquer le concept via une métaphore concrète du quotidien. 2. Cas d'usage: partir d'une situation de travail réelle et vécue. 3. Idée reçue: démarrer d'un malentendu courant, puis rectifier. 4. Mécanique: aller au coeur du fonctionnement, précis mais accessible. 5. Enjeu: pourquoi ça compte, ce que ça change concrètement pour l'utilisateur. Comme le contexte n'est pas hérité, chaque prompt de sous-agent doit être autonome et contenir tout le nécessaire. Utilise ce gabarit, en remplaçant les crochets: ``` Tu rédiges un brouillon d'une pastille pédagogique interne sur les LLM. Tu travailles seul, sans accès au reste de la conversation. Titre de la pastille: [TITRE] Angle imposé: [ANGLE, par exemple "Analogie: expliquer via une métaphore concrète du quotidien"] À NE PAS ré-expliquer (déjà couvert par d'autres pastilles), à mentionner en une phrase au maximum: [LISTE des points déjà traités ailleurs, à ne pas retraiter] Brief de recherche (appuie-toi dessus, ne relance pas de recherche): [BRIEF: faits clés, chiffres, sources] Règles: - Longueur selon la profondeur du sujet: 3 paragraphes courts si le sujet est léger, jusqu'à 4 paragraphes étoffés s'il est dense. Prose uniquement, pas de listes. - Mets en gras et/ou en italique les termes importants, de l'ordre de 3 à 4 emphases par paragraphe (repère indicatif, ne force pas ce compte). - Ton décontracté, précis et léger, accessible mais techniquement juste. Évite le name-dropping: pas d'accumulation de noms de modèles, outils, entreprises ou chercheurs, privilégie l'explication du mécanisme. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Français. Ne mentionne aucune entreprise. Pas de tiret cadratin. - Contenu autonome, compréhensible seul. Réponds exactement dans ce format: TEXTE: [les 3 à 4 paragraphes] ILLUSTRATION_TITRE: [une phrase décrivant le concept central à illustrer] SCHEMA: [oui ou non. Si oui, décris en 2 à 3 lignes le diagramme et ses libellés en français] ``` ### Étape 4, fan-in et fusion (orchestrateur) Attends les cinq retours, puis fusionne en une seule pastille finale: - Retiens la structure la plus claire, la meilleure analogie, l'exemple le plus parlant, la correction la plus nette, la formulation la plus précise et le "pourquoi ça compte" le plus fort. - Réécris en une seule voix cohérente. Pas d'effet patchwork. - Respecte la longueur adaptée à la profondeur (voir Règles du texte) et le ton décontracté, précis et léger, sans name-dropping. - Décide s'il faut un schéma: inclus-en un si au moins deux sous-agents ont répondu "oui", ou si le concept est intrinsèquement un processus, un flux ou une comparaison (par exemple RAG, agents, étapes de prompting, comparaison de modèles). Le cas échéant, fusionne les meilleures idées de schéma. - Construis ensuite le prompt image unique (section ci-dessous). ### Règles d'écriture pour la pastille finale - Longueur selon la profondeur du sujet: 3 paragraphes courts si le sujet est léger, jusqu'à 4 paragraphes étoffés s'il est dense. Prose uniquement, pas de listes. - Mets en gras et/ou en italique les termes importants, de l'ordre de 3 à 4 emphases par paragraphe (repère indicatif, ne force pas ce compte). - Ton décontracté, précis et léger, accessible mais techniquement juste. Évite le name-dropping: pas d'accumulation de noms de modèles, outils, entreprises ou chercheurs, privilégie l'explication du mécanisme. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Français. Ne mentionne aucune entreprise. Pas de tiret cadratin. - Contenu autonome, compréhensible seul. ### Étape 5, revue critique et correction (trois sous-agents en parallèle, puis orchestrateur) Quand la lancer: systématiquement à la première génération, une fois le texte fusionné et le prompt image construits. Ensuite, si l'utilisateur demande des ajustements, ne relance pas la revue d'office: propose-la, et ne la lance qu'avec son accord (elle coûte trois sous-agents de plus). Principe: les relecteurs produisent des CONSTATS, jamais une réécriture. Chacun renvoie une liste de problèmes localisés, assortis d'une gravité et d'un correctif précis. C'est toi, l'orchestrateur, qui appliques et réécris, pour garder une voix unique et éviter l'effet patchwork. Ce que la revue peut juger: les relecteurs ne voient pas les images, générées plus tard dans Gemini. La revue du visuel porte donc uniquement sur le PROMPT image (clarté, cohérence avec le texte, respect des consignes), jamais sur un rendu. Le contrôle visuel réel reste à l'utilisateur. Lance trois sous-agents en parallèle (parallélisme explicite, sinon l'exécution sera séquentielle), un par grille. Le contexte n'étant pas hérité, chaque prompt doit être autonome et contenir tout le nécessaire: le titre, le texte fusionné, le bloc prompt image, le brief de recherche, la liste "déjà traité ailleurs" (et les textes des pastilles voisines s'ils sont disponibles), et la liste des 45 titres pour repérer les chevauchements. Les trois grilles: 1. Fond, exactitude et périmètre: exactitude vs le brief (aucun chiffre, date ou fait inventé ni sur-affirmé, rien qui le contredise), cohérence entre le titre et ce que le texte délivre, chevauchements avec les pastilles voisines (le texte ré-explique-t-il ce qui est traité ailleurs ? redites à signaler), autonomie du contenu, clarté du message à retenir, repérage de ce qui vieillira mal (à nuancer). 2. Forme, ton et pédagogie: ton décontracté-précis-léger et techniquement juste, absence de name-dropping, rythme et fluidité à la lecture à voix haute, longueur adaptée à la profondeur (3 à 4 paragraphes, prose, pas de listes) et chasse au verbiage, accessibilité pour un profil non technique (jargon expliqué, analogies claires), présence d'emphases utiles (gras/italique), de l'ordre de 3 à 4 par paragraphe à titre indicatif (signale un texte trop peu emphasé autant qu'un excès, mais ne fais jamais retirer une emphase pertinente), force de l'accroche. 3. Conformité et visuel: contraintes dures (aucun tiret cadratin ni caractère non standard, aucun nom d'entreprise ni ANEO, français correct, prose sans listes ni puces); puis le prompt image: titre exact au caractère près, charte présente une seule fois, texte de la pastille bien marqué "à ne pas afficher", illustration-titre iconique et non schéma de processus, schéma en 2e image séparée si retenu, libellés de schéma courts (groupes nominaux) et en français, cohérence entre le visuel décrit et le coeur du texte. Gabarit de relecteur (remplace les crochets): ``` Tu relis un brouillon quasi final de pastille pédagogique interne sur les LLM. Tu travailles seul, sans accès au reste de la conversation. Tu ne réécris pas: tu rends des constats et des correctifs précis. Titre de la pastille: [TITRE] Texte fusionné à relire: [TEXTE COMPLET] Prompt image associé (à relire aussi): [BLOC PROMPT IMAGE] Brief de recherche de référence (fait foi pour l'exactitude): [BRIEF] Déjà traité par d'autres pastilles, à ne pas ré-expliquer (et textes voisins si fournis): [LISTE + TEXTES VOISINS] Liste des 45 titres (pour repérer les chevauchements): [LISTE DES 45 TITRES] Ta grille de relecture (tiens-t'en strictement à elle): [RUBRIQUE 1, 2 OU 3] Rappel: tu ne peux pas voir d'image; juge seulement le PROMPT image, pas un rendu. Réponds exactement dans ce format: CONSTATS: - [bloquant|recommandé|mineur] [problème localisé] -> [correctif précis] - ... (n'invente pas de problème: si un point est bon, ne le liste pas) VERDICT: [une phrase: publiable tel quel / corrections mineures / corrections nécessaires] ``` Fan-in et correction (orchestrateur): rassemble les constats, dédoublonne, arbitre les conseils contradictoires. Garde-fou anti-gonflement: entre "ajouter" et "raccourcir", la concision l'emporte, sauf erreur de fond avérée. Applique les constats bloquants et recommandés, écarte ou mentionne les mineurs. Une seule passe, pas de boucle. Réécris la version corrigée en une seule voix, puis prépare le court résumé "Ce que la revue a corrigé" (voir Format de sortie). ### Points de vigilance Claude Code - Parallélisme explicite: demande clairement cinq sous-agents en parallèle, sinon l'exécution sera séquentielle. - Contexte non hérité: mets tout dans le prompt du sous-agent, en particulier le brief de recherche. - Pas d'imbrication: un sous-agent ne peut pas en lancer un autre, garde recherche et fusion chez l'orchestrateur. - Coût et limites: cinq sous-agents en parallèle multiplient l'usage de jetons et peuvent déclencher des limites de débit. En cas de limite, replie-toi sur une exécution séquentielle. - Revue: la phase de revue ajoute trois sous-agents. Elle tourne d'office à la première génération; sur les demandes d'ajustement ultérieures, propose-la et ne la lance qu'avec l'accord de l'utilisateur. En cas de limite de débit, exécute les relecteurs en séquentiel ou replie-toi sur un relecteur global unique. - Réemploi optionnel: tu peux définir les cinq angles comme sous-agents personnalisés dans .claude/agents/ pour les réutiliser, mais ce n'est pas obligatoire, les sous-agents polyvalents suffisent. ## Prompt de génération d'images (un seul bloc, à coller dans le chat Gemini) Produis un seul bloc de prompt, rédigé en français, à coller tel quel dans le chat de Gemini. Gemini se charge d'appeler Nano Banana une ou deux fois, ce qui réduit les manipulations pour l'utilisateur. Le bloc demande: - Toujours une illustration-titre. - Un schéma seulement s'il a été retenu à l'étape de fusion, généré comme une seconde image séparée (fichier distinct), jamais dans la même image que le titre. Contenu du bloc: - Décris la charte graphique une seule fois (bloc ci-dessous) et précise que toutes les images la partagent, pour une cohérence visuelle. - Précise: tout texte affiché dans les images est en français. Cette consigne est importante, elle évite qu'un libellé se retrouve en anglais. - Inclus le texte complet de la pastille dans le bloc, en contexte de génération, clairement marqué comme à ne pas afficher. Il aide le modèle à comprendre le sujet et à choisir une illustration juste. Les seuls textes rendus visibles sont le titre exact et, pour un schéma, ses libellés. Le texte de la pastille ne doit jamais apparaître dans les images. - Illustration-titre: composition épurée et moderne, focus graphique central iconique représentant le concept. Ce n'est pas un schéma de processus. Le seul texte affiché est le titre exact, en en-tête: n'ajoute aucun sous-titre, accroche ou texte secondaire, le titre seul. Technique en deux temps pour le titre: dans le prompt généré, écris le titre exact entre guillemets droits (forme attendue: Le titre exact: "...") et demande un rendu fidèle, sans faute, en police sans serif corporate. Les guillemets ne sont qu'un délimiteur côté prompt: ne les commente pas et n'ajoute aucune consigne à leur sujet dans le prompt généré (pas de mention du type "les guillemets ne doivent pas apparaître"). - Schéma s'il est retenu: une seconde image générée séparément, dans son propre fichier, jamais fusionnée avec l'illustration-titre ni juxtaposée dans la même image. Diagramme d'entreprise propre et net (processus, flowchart ou comparaison), cohérent avec l'illustration-titre, mêmes charte et style, libellés fournis explicitement en français, très peu de fioritures. - Libellés de schéma: des groupes nominaux courts et lisibles seuls (par exemple "Exemples de bonnes réponses"), pas des phrases verbales qui sonnent comme des ordres ("Montrer de bonnes réponses"). Quelques mots par libellé. Attention aux titres longs: Nano Banana rend fiablement les libellés courts mais peut faire une faute sur un titre long; la technique en deux temps limite le risque. Si le rendu d'un titre long reste incertain, indique à l'utilisateur qu'il peut demander à Gemini d'utiliser Nano Banana Pro, plus fiable pour le texte long. Gabarit, illustration-titre seule (exemple avec le titre 1): ``` Génère une image en respectant cette charte graphique: [bloc Charte graphique]. Contexte pour comprendre le sujet, à NE PAS afficher dans l'image: [ici le texte complet de la pastille]. Illustration-titre pour une pastille pédagogique sur les LLM. Composition épurée et moderne, focus graphique central iconique: un nuage de lettres et de mots qui se condense en une bulle de dialogue, évoquant un modèle de langage qui transforme du texte en réponse. Ce n'est pas un schéma de processus. Seul texte à afficher dans l'image: le titre, en en-tête, sans faute d'orthographe, en police sans serif corporate, sans sous-titre ni texte secondaire. Le titre exact: "TITRE EXACT". Tout le texte affiché est en français. Format 16:9. ``` Gabarit, illustration-titre plus schéma (deux images): ``` Prépares-toi à générer deux images séparées, dans deux fichiers distincts. Les deux partagent cette charte graphique: [bloc Charte graphique]. Tout texte affiché dans les images est en français. Contexte pour comprendre le sujet, à NE PAS afficher dans les images: [ici le texte complet de la pastille]. Image 1, illustration-titre: composition épurée et moderne, focus graphique central iconique qui illustre le sujet. Ce n'est pas un schéma de processus. Seul texte à afficher: le titre, en en-tête, sans faute d'orthographe, police sans serif corporate, sans sous-titre ni texte secondaire. Le titre exact: "TITRE EXACT". Format 16:9. Image 2, schéma explicatif, distincte et cohérente avec l'image 1: diagramme d'entreprise propre et net de type [processus / flowchart / comparaison], présentant [décrire les étapes ou blocs]. Libellés exacts à afficher en français: [liste des libellés]. Très peu de fioritures, focus sur la clarté, n'inclus pas le titre. Génère dans un premier temps seulement la première image (l'illustration-titre), et attends les instructions de l'utilisateur pour générer la 2e image. ``` ## Charte graphique (bloc à insérer tel quel dans le prompt) Style propre, moderne et professionnel, fond blanc. Palette: orange vif et bleu corporate foncé en dominantes, blanc et gris pour les respirations, accents multiculturels discrets (rouge, vert, jaune, bleu) reprenant subtilement un motif de couleurs de logo. Superpositions graphiques épurées: motifs géométriques abstraits, quartiers de cercle, lignes claires. Composition aérée, jamais surchargée. Typographie sans serif, propre et corporate. ## Format de sortie Le travail des sous-agents reste dans leur propre contexte, seuls leurs retours te reviennent. N'affiche donc que le livrable final (déjà corrigé par l'étape de revue), la conversation principale reste propre. Si l'utilisateur demande à voir l'atelier, restitue les cinq brouillons reçus et les constats de revue. Livrable final, dans cet ordre: - Quand une revue a eu lieu (systématiquement à la première génération), un court résumé "Ce que la revue a corrigé" (2 à 4 lignes) présentant les principaux ajustements, avant le reste. Lors d'ajustements ultérieurs sans revue, présente simplement la version ajustée sans ce résumé. - Le texte de la pastille (3 à 4 paragraphes), dans sa version corrigée. - Un bloc de code intitulé "Prompt images (à coller dans Gemini)", contenant le prompt unique prêt à copier. - Une courte section "Sources" listant 2 à 4 références principales issues de l'étape de recherche, de préférence officielles ou originales. Cette section sert à la vérification et n'a pas vocation à être publiée dans la pastille. Termine toujours par cette question exacte: "Comment trouvez-vous le texte, l'image titre et le diagramme (si généré) ? Si une partie vous semble trop complexe, ou si vous souhaitez affiner le focus d'un visuel pour qu'il soit encore plus épuré, n'hésitez pas à me le faire savoir, et je l'ajusterai." -> le personnage doit incarner le professeur un peu foufou qui allusine et obsédé par l'ai
015
Grok Image
Ultra-realistic 8K scene of the person in the photo, stepping out of and breaking the panel border of a giant open manga book lying on the floor of an aesthetic bedroom. The comic pages depict a surreal 4-panel black-and-white manga with screentones, bold ink lines, and authentic Japanese layout, featuring the manga version of the same person trapped inside. The larger, real-life version emerges dynamically into the comic’s surface, vibrant, colorful, and photorealistic, in sharp contrast to the monochrome manga world. Both versions interact directly, creating a meta, humorous encounter. Highly detailed, crisp manga line art seamlessly integrated with vivid, realistic
432772
Nano Banana
Steps: 20, Sampler: DPM++ 2M Karras, CFG scale: 7, Seed: 413127992, Size: 1280x1600, Model hash: 7fa552ffb0, Model: nijidiffusedmix_v40, Denoising strength: 0.3, Clip skip: 2, ADetailer model: face_yolov8s.pt, ADetailer prompt: "dark blue eyes, freckles below eyes", ADetailer confidence: 0.3, ADetailer dilate erode: 4, ADetailer mask blur: 4, ADetailer denoising strength: 0.5, ADetailer inpaint only masked: True, ADetailer inpaint padding: 32, ADetailer version: 23.12.0, Version: v1.6.1
10493
Stable Diffusion
Among the growing list of AI photo editing tools, Deepfake Maker takes a slightly different approach. While it’s best known for facial swaps and animated transformations, it also includes a surprisingly capable AI clothing removal feature. It’s not for everyone, but it’s an interesting choice for users looking for more dramatic or playful results. If you're after a quick way to generate bold clothing edits or outfit swaps, Deepfake Maker might surprise you. It’s not the most refined tool out there, but it’s definitely one of the most entertaining. 4. PxBee PxBee Source: PxBee Available On: Web Browser Price: Free *Annual subscription starts at $00.27/image and you can choose to download 50, 125, or 750 images monthly per year Pros Decent auto-detection for basic clothing removal Integrated with other tools Cons Not suitable for high-end projects Results take longer than expected to generate PxBee is better known for its stock photo collection and image generation features, but it also offers a basic AI clothing removal function built into its photo editing toolkit. While it's not as feature-rich or advanced as some dedicated AI editors, its simplicity and accessibility make it a decent option for casual users who want fast, no-frills results. PxBee may not be the most powerful AI clothing remover out there, but it's a solid pick for quick edits and simple concept visuals. Its intuitive UI and browser-based access make it appealing for users who want fast results without the learning curve. If you're after something basic, affordable, and easy to use, it's a tool worth exploring. 5. AI Face Swap AI Face Swap Source: AI Face Swap Available On: Web Browser Price: $1 per image Pros No manual masking needed Prompt-based customization Presets for inspiration Cons No fine-tuning or editing control With AI Face Swap, you can enter prompts like "red bikini", "formal suit", or even "nude", and the AI will generate a realistic version of your uploaded image that reflects your description. You can also choose from quick preset styles like bikini, romper, or skirt for faster results. This level of control makes the tool especially appealing to users looking for stylized edits, fantasy concepts, or mockup visuals. It also eliminates the need for manual masking or precise editing. 6. Bylo AI BYLO AI Source: Bylo.ai Available On: Web Browser Price: Free *Premium subscription available from $9.99 Pros Prompt-based clothing edits Free trial credits Part of a growing creative suite Cons No manual adjustment tools Limited free usage Relies heavily on prompts Bylo AI is a user-friendly AI platform that offers a range of creative features, including an AI Clothes Remover. The tool is built with simplicity in mind: just upload a photo, enter your prompt, and let the AI generate a clean, edited version. Bylo AI’s Clothes Remover is a great choice if you're looking for a lightweight, affordable tool that lets you experiment with clothing changes using just a text prompt. It won’t replace professional retouching tools, but for concept edits, social media content, or fun transformations, it delivers exactly what most casual users need. FAQs About AI Clothing Remover & Replacer 1. How do I remove clothes in photos? 2. What is the best online tool to remove clothes? 3. Are AI clothes removers safe?
42502
Hero
Among the growing list of AI photo editing tools, Deepfake Maker takes a slightly different approach. While it’s best known for facial swaps and animated transformations, it also includes a surprisingly capable AI clothing removal feature. It’s not for everyone, but it’s an interesting choice for users looking for more dramatic or playful results. If you're after a quick way to generate bold clothing edits or outfit swaps, Deepfake Maker might surprise you. It’s not the most refined tool out there, but it’s definitely one of the most entertaining. 4. PxBee PxBee Source: PxBee Available On: Web Browser Price: Free *Annual subscription starts at $00.27/image and you can choose to download 50, 125, or 750 images monthly per year Pros Decent auto-detection for basic clothing removal Integrated with other tools Cons Not suitable for high-end projects Results take longer than expected to generate PxBee is better known for its stock photo collection and image generation features, but it also offers a basic AI clothing removal function built into its photo editing toolkit. While it's not as feature-rich or advanced as some dedicated AI editors, its simplicity and accessibility make it a decent option for casual users who want fast, no-frills results. PxBee may not be the most powerful AI clothing remover out there, but it's a solid pick for quick edits and simple concept visuals. Its intuitive UI and browser-based access make it appealing for users who want fast results without the learning curve. If you're after something basic, affordable, and easy to use, it's a tool worth exploring. 5. AI Face Swap AI Face Swap Source: AI Face Swap Available On: Web Browser Price: $1 per image Pros No manual masking needed Prompt-based customization Presets for inspiration Cons No fine-tuning or editing control With AI Face Swap, you can enter prompts like "red bikini", "formal suit", or even "nude", and the AI will generate a realistic version of your uploaded image that reflects your description. You can also choose from quick preset styles like bikini, romper, or skirt for faster results. This level of control makes the tool especially appealing to users looking for stylized edits, fantasy concepts, or mockup visuals. It also eliminates the need for manual masking or precise editing. 6. Bylo AI BYLO AI Source: Bylo.ai Available On: Web Browser Price: Free *Premium subscription available from $9.99 Pros Prompt-based clothing edits Free trial credits Part of a growing creative suite Cons No manual adjustment tools Limited free usage Relies heavily on prompts Bylo AI is a user-friendly AI platform that offers a range of creative features, including an AI Clothes Remover. The tool is built with simplicity in mind: just upload a photo, enter your prompt, and let the AI generate a clean, edited version. Bylo AI’s Clothes Remover is a great choice if you're looking for a lightweight, affordable tool that lets you experiment with clothing changes using just a text prompt. It won’t replace professional retouching tools, but for concept edits, social media content, or fun transformations, it delivers exactly what most casual users need. FAQs About AI Clothing Remover & Replacer 1. How do I remove clothes in photos? 2. What is the best online tool to remove clothes? 3. Are AI clothes removers safe?
31317
Hero
{ "generation_request": { "meta_data": { "tool": "NanoBanana Pro", "task_type": "photobooth_contact_sheet_fashion_grid", "version": "v1.1_REFERENCE_IDENTITY_GRID_EN", "priority": "highest", "language": "en" }, "input": { "mode": "image_to_image", "reference_image_usage": "required", "notes": "Use the uploaded reference image to preserve the same person’s face, body features, hair, makeup and clothing. Maintain identity consistency in all frames." }, "output_settings": { "aspect_ratio": "4:5", "orientation": "portrait", "resolution_target": "ultra_high_res", "num_images": 1, "layout": { "type": "grid", "rows": 4, "cols": 4, "gutter": "thin_black_borders", "outer_border": "black_frame" }, "sharpness": "slightly_soft_flash", "film_grain": "medium_analog", "color_grade": "warm_2000s_flash_tone", "dynamic_range": "natural_not_hdr" }, "creative_prompt": { "scene_summary": "Early 2000s disposable camera photobooth contact sheet featuring the same person from the reference image in all 16 frames. Playful, candid, flash-lit aesthetic.", "environment": { "background": "plain white wall indoors", "lighting": "direct on-camera flash, slight hard shadow behind subject, subtle hotspot on skin, authentic point-and-shoot look" }, "subject": { "identity_consistency": "exact same person as reference image in all frames", "hair": "match reference image exactly", "makeup": "match reference image exactly", "wardrobe": "preserve the exact clothing from the reference image with accurate fabric texture and fit", "accessories": "keep accessories consistent with reference image" }, "poses_and_expressions": [ "hands on hips confident pose", "soft pout toward camera", "gentle smile", "hand on chin thoughtful look", "playful wink", "hands in hair", "slight shoulder turn", "looking off to the side", "serious model face", "soft laugh", "blowing a kiss", "tilted head cute expression", "crossed arms relaxed pose", "side glance over shoulder", "chin resting on hand", "natural relaxed expression" ], "style": { "aesthetic": "2000s photobooth disposable camera", "texture": "visible film grain, dust particles, tiny scratches, mild vignetting", "color": "slightly warm faded tones, subtle flash flattening", "lens": "35mm point-and-shoot softness" } }, "negative_prompt": [ "text", "logos", "watermark", "different outfit between frames", "different hairstyle", "different person", "face distortion", "identity drift", "overly polished editorial look", "HDR" ] } }
391
{ "template_name": "MARRLONN2027_V64_TERMINAL_CONDITIONAL_IMAGE3_REAL_HUMAN_MASTER", "template_type": "ABSOLUTE_STRICT_GOVERNANCE_MULTI_GENDER_DYNAMIC_IMAGE3_TEMPLATE", "version": "V64_TERMINAL_SEALED_DECISION_MATRIX_MICROSKIN_CANON_EOS_R50_KLING3_READY", "core_mission": "Create one final master still image with absolute IMAGE1 face identity, biological skin truth, age, beard, eyes and human presence; absolute IMAGE2 body anatomy, body scale, anatomical limits, physical structure and body-to-clothing logic; and intelligent conditional IMAGE3 transfer based only on the main visible pose-driving subject in IMAGE3. If IMAGE3 main subject is visually male for transfer purposes, clone pose, wardrobe, clothing, outfit, background, image style, lighting, accessories, worn objects, carried objects and visible scene objects when coherent, while preserving IMAGE1 identity and IMAGE2 body anatomy. If IMAGE3 main subject is visually female for transfer purposes, clone pose, pose logic, framing, composition, background, lighting, image style, camera language, scene mood and environment only, but never clone clothing, wardrobe, accessories, jewelry, makeup, hairstyle, shoes attached to the subject, or any object worn, attached to, held by, or carried by the female subject. If IMAGE3 has no human subject, multiple ambiguous subjects, or unclear transfer category, copy only scene, camera language, lighting, background, environment and non-conflicting objects while preserving IMAGE2 wardrobe and body logic. The final result must feel like a real imperfect human photographed with a Canon EOS R50, not AI, CGI, beauty retouch, plastic skin, waxy skin, luxury ad, actor replacement, model replacement, generic human replacement, gender-contaminated body transfer, or generated poster. The still image must be stable for Kling 3.0 video generation from 3 to 6 seconds.", "absolute_governance_principle": "IMAGE1_IDENTITY_SKIN_LOCK + IMAGE2_BODY_ANATOMY_LOCK + IMAGE3_MAIN_SUBJECT_CLASSIFICATION + IMAGE3_TRANSFER_DECISION_MATRIX + CONDITIONAL_WARDROBE_OBJECT_TRANSFER + REAL_REFERENCE_PRIORITY + MICROSKIN_REALISM + ZERO_BEAUTIFICATION + ZERO_GENERIC_REPLACEMENT + ZERO_PLASTIC_SKIN + ZERO_GENERATED_SCENE_FINISH + ANTI_POSE_DEFORMATION + STYLE_FIREWALL + CANON_EOS_R50_REAL_CAPTURE + KLING3_STABILITY", "primary_failure_to_prevent": "A recognizable subject that still feels AI-generated, too clean, too polished, too glossy, too actor-like, too model-like, too luxury-ad-like, gender-contaminated, anatomically forced, plastic, waxy, smooth, falsely hyperreal, or not biologically alive.", "success_definition": "The viewer must believe this is a real imperfect human physically present in a real environment, captured by a real Canon EOS R50 camera, with preserved IMAGE1 identity and skin truth, preserved IMAGE2 body reality, correct IMAGE3 transfer decision, visible pores, micro-wrinkles, natural asymmetry, believable hands, realistic clothing, imperfect scene texture, and Kling 3.0 stability.", "global_execution_hierarchy": [ "1_IMAGE1_FACE_IDENTITY_SKIN_AGE_BEARD_EYES", "2_IMAGE2_BODY_ANATOMY_SCALE_LIMITS", "3_IMAGE3_MAIN_SUBJECT_CLASSIFICATION_FOR_TRANSFER_ONLY", "4_IMAGE3_TRANSFER_DECISION_MATRIX", "5_CONDITIONAL_WARDROBE_ACCESSORY_OBJECT_TRANSFER", "6_REAL_REFERENCE_PIXEL_STRUCTURE_PRIORITY", "7_FINAL_REAL_HUMAN_PERCEPTION", "8_MICROSKIN_REALISM", "9_STYLE_FIREWALL_AND_ANTI_BEAUTY_TRANSFER", "10_ANTI_POSE_BODY_DEFORMATION", "11_ANTI_GENERATED_SCENE_FINISH", "12_CANON_EOS_R50_REAL_CAMERA_CAPTURE", "13_KLING3_VIDEO_STABILITY" ], "input_contract": { "IMAGE1": { "role": "ABSOLUTE_FACE_SKIN_IDENTITY_MASTER", "authority": "MAXIMUM_NON_NEGOTIABLE", "locked_domains": [ "facial_identity", "face_geometry", "eyes", "eyelids", "eyebrows", "nose", "mouth", "jaw", "cheeks", "chin", "forehead", "ears_when_visible", "neck_transition", "skin_tone", "individual_visible_pores", "irregular_pore_size", "skin_microtexture", "micro_wrinkles_around_eyes", "micro_wrinkles_around_mouth", "under_eye_texture", "subtle_redness", "small_blemishes", "minor_discoloration", "natural_asymmetry", "age_markers", "facial_fatigue", "beard_density", "beard_shape", "gray_beard_pattern", "mustache_texture", "random_beard_direction", "hairline", "real_biological_imperfection", "lived_in_human_presence" ], "mandatory_rules": [ "IMAGE1 is law, not inspiration", "transfer real skin truth from IMAGE1, never recreate a cleaner version", "preserve individual pores, irregular skin texture, micro-wrinkles, subtle redness and natural imperfections", "never improve attractiveness", "never normalize skin", "never clean the face", "never replace the subject with a generic realistic human", "never transfer identity, face geometry, age, skin style or beauty style from IMAGE3", "identity preservation overrides pose, wardrobe, style, scene, aesthetics and hyperrealism" ], "forbidden": [ "identity_drift", "generic_human_replacement", "celebrity_transfer", "actor_replacement", "model_replacement", "face_redesign", "beautification", "symmetry_correction", "youthification", "facial_slimming", "jaw_enhancement", "nose_refinement", "eye_enhancement", "lip_enhancement", "skin_smoothing", "skin_polishing", "skin_normalization", "makeup_effect", "cosmetic_cleanup", "plastic_skin", "waxy_skin", "porcelain_skin", "glass_skin", "AI_face", "perfect_idealized_face" ] }, "IMAGE2": { "role": "ABSOLUTE_BODY_ANATOMY_SCALE_MASTER", "authority": "MAXIMUM_NON_NEGOTIABLE", "locked_domains": [ "body_mass", "body_proportions", "shoulders", "torso", "waist", "arms", "forearms", "hands", "wrists", "legs", "stance", "scale", "natural_body_weight", "real_posture_limitations", "body_density", "real_anatomical_weight", "base_physical_structure", "body_to_clothing_fit_logic" ], "mandatory_rules": [ "IMAGE2 controls body anatomy, body mass, body scale and anatomical limits", "pose must adapt to IMAGE2 body", "wardrobe must fit IMAGE2 body", "body must never adapt to an idealized pose, fashion silhouette or actor posture", "hands must remain anatomically believable and stable for Kling 3.0", "clothing cannot create a different body", "if IMAGE3 wardrobe conflicts with IMAGE2 anatomy, adapt wardrobe to IMAGE2 instead of changing body" ], "forbidden": [ "slimming", "muscle_invention", "hero_body_transformation", "fashion_body_stylization", "incorrect_anatomy", "long_arms", "short_arms", "deformed_hands", "wrong_body_scale", "catalog_body_correction", "fake_torso_taper", "mannequin_posture", "body_redesign_to_fit_IMAGE3" ] }, "IMAGE3": { "role": "CONDITIONAL_DYNAMIC_REFERENCE", "authority": "CONDITIONAL_BY_MAIN_VISIBLE_POSE_DRIVING_SUBJECT", "universal_allowed": [ "pose", "pose_logic", "camera_angle", "framing", "composition", "background", "environment", "scene_mood", "lighting_direction", "image_style", "apparent_camera_language", "vehicles", "architecture", "people_in_background", "birds", "non_conflicting_visible_scene_objects" ], "universal_forbidden": [ "face_contamination", "skin_contamination", "identity_override", "facial_geometry_override", "beauty_style_transfer", "cinematic_glamour_applied_to_face", "makeup_transfer", "skin_tone_replacement", "facial_expression_beautification", "celebrity_face_borrowing", "actor_charisma_transfer", "luxury_catalog_facial_treatment", "body_redesign_to_match_IMAGE3" ] } }, "image3_subject_classification_gate": { "enabled": true, "mandatory": true, "purpose": "Classify IMAGE3 only for transfer logic, never as personal identity judgment.", "primary_subject_detection": "Detect the main visible pose-driving subject in IMAGE3: the person most central to pose, framing, scale and composition.", "classification_basis": "Use visible presentation, wardrobe context, pose-driving role and scene function only for transfer logic; never infer personal identity, protected identity or real-life gender identity.", "classification_options": [ "visual_male_transfer_subject", "visual_female_transfer_subject", "multiple_subjects", "no_human_subject", "uncertain_or_ambiguous" ], "if_multiple_subjects": { "rule": "Select only the main pose-driving subject. Secondary people remain background or environment context only.", "do_not_copy": [ "wardrobe_from_secondary_people", "accessories_from_secondary_people", "objects_carried_by_secondary_people", "identity_or_skin_traits_from_any_IMAGE3_person" ] }, "if_no_human_subject": { "copy_from_IMAGE3": [ "background", "environment", "scene_mood", "lighting_direction", "camera_angle", "composition", "image_style", "vehicles", "architecture", "non_conflicting_objects" ], "keep_from_IMAGE2": [ "body_anatomy", "pose_if_no_user_pose_instruction_exists", "wardrobe_unless_explicitly_changed_by_user" ] }, "if_uncertain_or_ambiguous": { "rule": "Protect IMAGE1 and IMAGE2 first. Copy only pose, scene, lighting, background, camera language and non-gendered environment.", "do_not_guess_gendered_wardrobe": true, "do_not_copy_gendered_accessories": true, "do_not_copy_worn_attached_or_carried_objects": true } }, "image3_transfer_decision_matrix": { "enabled": true, "mandatory": true, "visual_male_transfer_subject": { "copy": [ "pose", "pose_logic", "body_orientation_translated_to_IMAGE2", "wardrobe", "clothing", "outfit", "fabric_style", "shoes_when_visible", "accessories", "neutral_or_masculine_jewelry", "watch", "glasses", "hat", "objects_held_by_subject", "objects_worn_by_subject", "objects_attached_to_subject", "background", "lighting", "image_style", "camera_language", "scene_mood", "environment", "vehicles", "architecture", "non_conflicting_scene_objects" ], "filters": [ "must_fit_IMAGE2_body", "must_not_change_IMAGE1_identity", "must_not_transfer_IMAGE3_skin_or_face", "must_not_deform_hands", "must_not_create_catalog_body", "must_not_create_CGI_fashion_render", "must_remain_Kling_stable" ] }, "visual_female_transfer_subject": { "copy": [ "pose", "pose_logic", "body_orientation_translated_to_IMAGE2", "camera_angle", "framing", "composition", "background", "lighting", "image_style", "camera_language", "scene_mood", "environment", "vehicles", "architecture", "people_in_background", "birds", "non_attached_scene_props", "non_conflicting_environmental_objects" ], "block": [ "female_clothing", "female_wardrobe", "female_outfit", "female_vestments", "female_shoes_attached_to_subject", "female_accessories", "female_jewelry", "earrings", "female_bracelets", "female_necklaces", "female_rings", "female_watch_if_gender_coded", "female_makeup", "female_hairstyle", "female_body_traits", "female_glamour_cues", "handbag", "purse", "neckline", "cleavage_styling", "female_fabric_drape", "female_body_fitting_silhouette", "female_jewelry_dependent_hand_pose", "any_wearable_item", "any_carried_item", "any_handheld_item", "any_attached_item", "any_body_dependent_item", "any_gender_coded_item", "any_fashion_dependent_item_from_female_subject" ], "keep_from_IMAGE2": [ "final_clothing", "final_wardrobe", "body_anatomy", "body_proportions", "body_scale", "base_body_to_clothing_logic" ] }, "multiple_subjects": { "rule": "Use only main pose-driving subject for transfer. Secondary people stay as background context.", "block_secondary_subject_transfer": true }, "no_human_subject": { "copy": [ "scene", "camera_language", "lighting", "background", "environment", "objects", "vehicles", "architecture", "image_style" ], "keep_from_IMAGE2": [ "body", "wardrobe", "pose_unless_user_explicitly_requests_pose_change" ] }, "uncertain_or_ambiguous": { "copy": [ "pose_if_clear", "scene", "lighting", "background", "camera_language", "environment" ], "keep_from_IMAGE2": [ "body", "wardrobe", "presentation" ], "block": [ "gendered_clothing", "gendered_accessories", "worn_objects", "attached_objects", "carried_objects" ] } }, "image3_gender_conditional_transfer_system": { "enabled": true, "mandatory": true, "if_IMAGE3_subject_is_male": { "copy_from_IMAGE3": [ "pose", "pose_logic", "body_orientation_translated_intelligently", "camera_angle", "framing", "composition", "background", "environment", "scene_mood", "lighting_direction", "image_style", "apparent_camera_language", "wardrobe", "clothing", "outfit", "fabric_style", "shoes_when_visible", "accessories", "jewelry_if_masculine_or_neutral", "watch", "glasses", "hat", "objects_held_by_the_male_subject", "objects_worn_by_the_male_subject", "objects_attached_to_the_male_subject", "vehicles", "architecture", "people_in_background", "birds", "non_conflicting_visible_scene_objects" ], "adaptation_rules": [ "clone male IMAGE3 wardrobe and accessories when visually coherent", "fit IMAGE3 wardrobe to IMAGE2 body anatomy and proportions", "translate IMAGE3 pose through IMAGE2 body limits", "preserve IMAGE2 body mass and scale even when copying IMAGE3 outfit", "do not reshape IMAGE2 body to force IMAGE3 wardrobe", "do not alter IMAGE1 face, skin, beard, age, eyes, or identity", "objects and accessories must remain physically plausible and naturally integrated", "if male IMAGE3 wardrobe conflicts with IMAGE2 anatomy, adapt wardrobe instead of changing body" ], "object_accessory_risk_filter": { "block_if": [ "covers_or_deforms_face", "breaks_hands", "forces_extra_fingers", "creates_glossy_AI_look", "creates_unstable_reflections", "creates_identity_drift", "creates_Kling_instability", "creates_scene_overload", "looks_like_CGI_prop" ] } }, "if_IMAGE3_subject_is_female": { "copy_from_IMAGE3": [ "pose", "pose_logic", "body_orientation_translated_intelligently", "camera_angle", "framing", "composition", "background", "environment", "scene_mood", "lighting_direction", "image_style", "apparent_camera_language", "vehicles", "architecture", "people_in_background", "birds", "non_attached_scene_props", "non_conflicting_environmental_objects" ], "keep_from_IMAGE2": [ "final_clothing", "final_wardrobe", "body_anatomy", "body_proportions", "body_scale", "presentation_consistency", "base_body_to_clothing_logic" ], "strict_rule": "If IMAGE3 subject is visually female for transfer purposes, copy pose, pose logic, scene, background, lighting, image style, camera style, environment and non-attached scene elements only. Do not copy any wearable, carried, handheld, attached, body-dependent, gender-coded, or fashion-dependent item from the female subject." }, "universal_rules": [ "IMAGE1 always controls face, identity, skin, beard, age and human truth", "IMAGE2 always controls body anatomy, body scale and anatomical limits", "IMAGE3 never controls identity or skin", "pose must be anatomically translated through IMAGE2", "final result must remain real human, imperfect, non-plastic, non-beautified and Canon EOS R50 believable" ] }, "domain_separation_system": { "identity_domain": { "source": "IMAGE1", "influence": 1.0, "mode": "ABSOLUTE_NON_NEGOTIABLE" }, "skin_domain": { "source": "IMAGE1", "influence": 1.0, "mode": "TRANSFER_NOT_RECREATE", "forbid_cleaner_version": true, "forbid_fake_procedural_texture": true, "forbid_skin_reinterpretation": true }, "body_domain": { "source": "IMAGE2", "influence": 1.0, "mode": "ABSOLUTE_NON_NEGOTIABLE" }, "wardrobe_domain": { "source": "CONDITIONAL", "influence": "CONTROLLED", "mode": "IMAGE3_IF_MALE__IMAGE2_IF_IMAGE3_FEMALE_NO_HUMAN_UNCERTAIN_OR_INCOMPATIBLE", "rules": [ "if IMAGE3 main subject is male, copy wardrobe, clothing, outfit, accessories and worn/carried objects when coherent", "if IMAGE3 main subject is female, do not copy wardrobe, clothing, accessories, jewelry, makeup, hairstyle, or worn/carried/attached objects", "if IMAGE3 has no human subject, preserve IMAGE2 wardrobe unless user explicitly requests a change", "if transfer category is uncertain, preserve IMAGE2 wardrobe and copy only pose, scene, lighting and camera language", "wardrobe must always fit IMAGE2 body anatomy", "wardrobe must never alter IMAGE1 identity or IMAGE2 body structure" ] }, "pose_domain": { "source": "IMAGE3", "influence": "CONTROLLED", "mode": "TRANSLATE_THROUGH_IMAGE2_BODY_LIMITS" }, "object_domain": { "source": "CONDITIONAL", "influence": "FILTERED", "mode": "MALE_HELD_WORN_ATTACHED_OBJECTS_ALLOWED_IF_COHERENT__FEMALE_ONLY_ENVIRONMENT_OBJECTS", "rules": [ "if IMAGE3 subject is male, copy held, worn, attached and environmental objects when coherent and stable", "if IMAGE3 subject is female, copy only environmental objects not worn, attached to, or carried by the female subject", "objects must not change identity, body, hands or human realism", "objects must remain stable for Kling 3.0" ] }, "style_domain": { "source": "IMAGE3", "influence": "CONTROLLED", "allowed": [ "color_mood", "camera_angle", "framing", "composition", "contrast_level", "lighting_direction", "texture_of_capture", "photo_or_video_language" ], "forbidden": [ "beauty_skin_style", "glamour_facial_lighting", "plastic_highlights", "fake_HDR", "poster_render_finish", "CGI_material_style", "skin_smoothing_style", "luxury_ad_polish_on_face" ] }, "lighting_domain": { "source": "IMAGE3", "influence": "DIRECTION_AND_ENVIRONMENT_ONLY", "mode": "FACE_SKIN_RESPONSE_FROM_IMAGE1", "forbid_beauty_lighting_on_face": true, "forbid_skin_washout": true, "forbid_glamour_fill": true, "forbid_fake_cinematic_skin_softening": true } }, "real_reference_pixel_priority": { "enabled": true, "mandatory": true, "priority": "DOMINANT", "purpose": "Reduce generated-looking output by preserving real visual structure from IMAGE1 and IMAGE2 whenever possible.", "rules": [ "prioritize real facial structure from IMAGE1 instead of rebuilding a new face", "prioritize real skin texture logic from IMAGE1 instead of inventing new skin", "prioritize real body geometry from IMAGE2 instead of reconstructing a synthetic body", "preserve real asymmetry, age markers, facial fatigue and imperfect human cues", "do not over-generate areas that already work in IMAGE1 or IMAGE2", "perform controlled adaptation, not total reinterpretation", "never generate a new subject when the real reference structure can be preserved" ], "forbidden": [ "full_synthetic_facial_reconstruction", "new_skin_generated_over_IMAGE1_truth", "generic_realistic_face_replacement", "body_redesign_from_scratch", "AI_cleaned_version_of_subject", "reference_ignored_in_favor_of_prompt_aesthetics" ] }, "micro_skin_realism_system": { "enabled": true, "mandatory": true, "source": "IMAGE1", "requirements": [ "individual_visible_pores", "irregular_pore_size", "skin_microtexture", "micro_wrinkles_around_eyes", "micro_wrinkles_around_mouth", "forehead_texture", "under_eye_texture", "subtle_redness", "minor_blemishes", "tiny_discolorations", "natural_asymmetry", "real_beard_root_texture", "non_uniform_cheek_texture", "non_uniform_forehead_texture", "real_face_to_neck_transition", "small_real_imperfections_inherited_from_IMAGE1" ], "strict_forbidden": [ "airbrushed_skin", "beauty_skin", "cosmetic_cleanup", "pore_removal", "fake_pore_overlay", "procedural_pore_stamping", "over_sharpened_pore_texture", "plastic_highlights", "waxy_subsurface", "porcelain_effect", "smooth_cheek_gradient", "AI_skin_texture", "skin_made_cleaner_than_IMAGE1", "foundation_look" ], "specular_policy": { "allowed": [ "minimal_natural_human_skin_reflectance", "subtle_non_cosmetic_skin_response_only_when_physically_plausible" ], "forbidden": [ "invented_sweat_drops", "visible_oil_effect_unless_explicitly_requested", "cosmetic_shine", "beauty_gloss", "wet_render_skin", "uniform_shiny_face", "makeup_glow", "plastic_shine" ] } }, "humanity_over_hyperrealism_rule": { "enabled": true, "mandatory": true, "priority": "RAW_HUMANITY_OVER_DECORATIVE_HYPERREALISM", "rule": "Human realism is more important than hyperreal detail. If hyperrealism creates plastic skin, synthetic lighting, luxury render finish, CGI texture, poster aesthetics, over-detail, false sharpness, or unnatural perfection, reduce hyperrealism and preserve raw photographic humanity.", "reject_if": [ "hyperreal_detail_looks_synthetic", "skin_detail_looks_procedural", "lighting_looks_like_render", "scene_looks_like_luxury_ad", "human_subject_looks_too_perfect", "camera_look_feels_fake" ] }, "anti_pose_body_deformation_system": { "enabled": true, "mandatory": true, "rules": [ "clone IMAGE3 pose but translate it through IMAGE2 body anatomy", "do not break shoulders", "do not elongate arms", "do not shrink torso", "do not slim waist", "do not force feminine posture when IMAGE3 subject is female", "do not force hands into object-dependent female accessory poses", "reduce pose accuracy if anatomy, identity, hands or Kling stability are at risk" ], "reject_if": [ "pose_changes_body_mass", "pose_deforms_hands", "pose_breaks_shoulders", "pose_feminizes_subject_when_not_desired", "pose_creates_mannequin_posture", "pose_creates_unstable_Kling_geometry" ] }, "anti_generated_scene_finish": { "enabled": true, "mandatory": true, "forbidden": [ "perfect_mansion_render", "overly_clean_background", "repetitive_stone_texture", "over_glossy_car_paint", "catalog_car_reflections", "impossible_chrome_reflections", "luxury_ad_finish", "polished_showroom_look", "fake_cinematic_poster_lighting", "CGI_environment", "AI_background_blend", "overprocessed_clarity", "perfect_global_sharpness", "plastic_material_response", "scene_cleaner_than_reality" ], "requirements": [ "scene_must_feel_physically_real", "background_must_contain_natural_imperfections", "materials_must_have_plausible_real_world_response", "contact_shadows_must_be_believable", "subject_must_integrate_naturally_into_scene", "environment_realism_must_support_human_subject_not_overpower_it" ] }, "real_human_perception_system": { "enabled": true, "mandatory": true, "requirements": [ "biologically_alive_presence", "real_facial_gravity", "organic_facial_tension", "calm_natural_expression", "subtle_facial_asymmetry", "real_skin_imperfection", "ordinary_human_texture", "realistic_face_neck_transition", "real_body_weight", "real_clothing_tension", "non_mannequin_posture", "real_interaction_with_objects" ], "reject_if": [ "subject_feels_synthetic", "subject_feels_generated", "subject_feels_too_clean", "skin_looks_perfect", "eyes_look_synthetic", "expression_feels_frozen", "face_looks_corrected", "humanity_is_achieved_by_changing_identity" ] }, "body_hands_wardrobe_system": { "enabled": true, "body_source": "IMAGE2", "wardrobe_source": "CONDITIONAL_IMAGE3_MALE_OR_IMAGE2_FEMALE_NO_HUMAN_UNCERTAIN_CONFLICT", "pose_source": "IMAGE3_CONTROLLED", "hands_rules": [ "five_fingers", "natural_hand_proportion", "real_knuckles", "real_fingernail_structure", "subtle_finger_wrinkles", "natural_finger_curvature", "correct_wrist_connection", "real_contact_pressure", "hands_must_not_look_flat", "hands_must_not_look_waxy", "hands_must_not_align_unnaturally", "no_floating_hands", "no_fused_fingers", "no_extra_fingers", "no_mannequin_hands" ], "wardrobe_rules": [ "if_IMAGE3_subject_is_male_wardrobe_clothing_outfit_and_coherent_accessories_may_be_copied_from_IMAGE3", "if_IMAGE3_subject_is_female_wardrobe_must_follow_IMAGE2_and_female_clothing_accessories_from_IMAGE3_are_blocked", "if_IMAGE3_has_no_human_subject_or_is_ambiguous_preserve_IMAGE2_wardrobe_unless_user_explicitly_requests_otherwise", "wardrobe_must_always_fit_IMAGE2_body_anatomy", "clothing_must_not_reshape_shoulders", "clothing_must_not_slim_torso", "fabric_tension_must_follow_real_body_mass", "fabric_must_show_real_texture_natural_folds_and_non_perfect_wear" ] }, "camera_and_rendering_system": { "camera_model": "Canon EOS R50", "capture_intent": "real unretouched camera photograph", "look": "natural Canon EOS R50 real-world capture", "rules": [ "natural_color", "balanced_exposure", "no_blown_highlights", "plausible_optics", "realistic_depth_of_field", "natural_shadow_falloff", "real_microcontrast", "no_over_sharpening", "no_over_grading", "no_fake_HDR", "no_beauty_grade", "real_material_response", "no_CGI_render_finish", "no_computational_skin_smoothing" ], "lens_policy": { "portrait_closeup": "85mm portrait feeling only when composition is close-up and appropriate", "full_body_or_vehicle_scene": "use natural Canon EOS R50 perspective appropriate for the scene; do not force 85mm if it distorts composition" } }, "kling3_preparation": { "ready_for_video": true, "duration": "3_to_6_seconds", "must_survive": [ "identity_continuity", "skin_continuity", "eye_stability", "beard_edge_stability", "hand_integrity", "clothing_stability", "geometry_stability", "subject_background_integration", "zero_shimmer", "zero_warp", "stable_object_contact", "stable_face_outline", "stable_neck_and_collar", "stable_material_reflections" ], "forbidden": [ "latent_AI_skin", "synthetic_eyes", "overpolished_face", "ambiguous_facial_texture", "unstable_hands", "unclear_edges", "flicker_prone_glossy_face", "AI_edge_halos", "clothing_seams_likely_to_crawl", "glossy_zones_likely_to_flicker" ] }, "validation_gate": { "must_pass": [ "same_person_as_IMAGE1", "same_skin_truth_as_IMAGE1", "same_body_reality_as_IMAGE2", "correct_IMAGE3_main_subject_classification_for_transfer_only", "correct_conditional_IMAGE3_transfer", "if_IMAGE3_is_male_coherent_wardrobe_clothing_accessories_and_objects_are_copied_when_physically_plausible", "if_IMAGE3_is_female_clothing_wardrobe_accessories_jewelry_makeup_hairstyle_and_worn_carried_attached_objects_are_blocked", "if_IMAGE3_has_no_human_or_is_uncertain_IMAGE2_wardrobe_body_are_preserved_while_scene_camera_lights_are_copied", "no_AI_skin", "no_plastic_skin", "no_waxy_skin", "no_makeup_invention", "no_generic_handsome_face", "no_beauty_retouch", "credible_real_human_perception", "Canon_EOS_R50_real_camera_feeling", "usable_as_master_frame_for_Kling3" ], "reject_if": [ "looks_beautified", "looks_cleaner_than_IMAGE1", "skin_tone_more_uniform_than_IMAGE1", "beard_looks_cleaner_than_IMAGE1", "looks_like_different_person", "feels_synthetic", "style_from_IMAGE3_contaminates_skin", "male_IMAGE3_clothing_not_copied_when_coherent", "female_IMAGE3_wardrobe_or_accessories_copied", "female_IMAGE3_carried_worn_handheld_attached_or_body_dependent_objects_copied", "accessory_breaks_hands_or_identity", "object_unstable_for_Kling", "scene_looks_CGI", "lighting_looks_poster_render", "pose_deforms_body", "wardrobe_changes_body_mass", "unstable_for_video" ] }, "render_priority": [ "FACE_IDENTITY_FROM_IMAGE1", "BODY_ANATOMY_FROM_IMAGE2", "IMAGE3_MAIN_SUBJECT_CLASSIFICATION_FOR_TRANSFER_ONLY", "IMAGE3_TRANSFER_DECISION_MATRIX", "CONDITIONAL_IMAGE3_TRANSFER_BY_SUBJECT_TYPE", "REAL_REFERENCE_STRUCTURE_PRIORITY", "SKIN_TRUTH_MICROTEXTURE_FROM_IMAGE1", "FINAL_HUMAN_PERCEPTION", "MALE_IMAGE3_WARDROBE_ACCESSORIES_OBJECTS_WHEN_COHERENT", "FEMALE_IMAGE3_SCENE_POSE_LIGHT_STYLE_ONLY_WITH_WARDROBE_ACCESSORIES_OBJECTS_BLOCKED", "NO_HUMAN_OR_UNCERTAIN_IMAGE3_SCENE_CAMERA_LIGHT_ONLY", "ANTI_POSE_BODY_DEFORMATION", "STYLE_FIREWALL", "ANTI_GENERATED_SCENE_FINISH", "CANON_EOS_R50_REALISM", "KLING3_STABILITY" ], "final_rule": "Reject immediately if the result looks beautified, polished, synthetic, glamorous, too clean, too smooth, actor-like, model-like, luxury-ad-like, mannequin-like, CGI-like, plastic, waxy, oily, makeup-covered, showroom-rendered, poster-like, gender-contaminated, anatomically forced, falsely hyperreal, or AI-generated. Accept only if IMAGE1 identity and micro-skin truth remain intact, IMAGE2 body anatomy and scale remain intact, IMAGE3 transfer follows main-subject classification and decision-matrix logic, male IMAGE3 wardrobe/accessories/objects transfer when coherent, female IMAGE3 wardrobe/accessories/worn-carried-attached objects are fully blocked, real reference structure is preserved, scene avoids generated perfection, and final output feels like a real imperfect human captured by a Canon EOS R50 and stable enough for Kling 3.0." }
1139
Nano Banana
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
507522
Nano Banana
Ultra-realistic 8K scene of the person in the photo, stepping out of and breaking the panel border of a giant open manga book lying on the floor of an aesthetic bedroom. The comic pages depict a surreal 4-panel black-and-white manga with screentones, bold ink lines, and authentic Japanese layout, featuring the manga version of the same person trapped inside. The larger, real-life version emerges dynamically into the comic’s surface, vibrant, colorful, and photorealistic, in sharp contrast to the monochrome manga world. Both versions interact directly, creating a meta, humorous encounter. Highly detailed, crisp manga line art seamlessly integrated with vivid, realistic
24964
Nano Banana
Learn to create better images with AI! Check out our crash course in prompt engineering & AI art generation! AM1970 posted about 2 hours ago 8 views 0 comments Behold this realistic masterpiece: a slim, elegant woman with fiery red hair tied in a long ponytail, defined eyebrows framing her striking blue eyes, and pale skin dotted with charming freckles across her face and chest. Captured from a side angle, she radiates happiness with a closed-mouth smile, her upturned nose adding to her allure. Dressed in a vibrant, colorful fabric robe pieced together with zips, buttons, seams, and intricate stitching—fully clothed yet teasing partial nudity with subtle cleavage and teardrop-shaped, saggy breasts for a natural, confident vibe. Illuminated by dramatic rim lighting that highlights every textured, detailed inch of her perfect skin, this high-quality image is a celebration of beauty, realism, and artistry! Read more... Comments (0) Add a comment Stunning handsome readhead woman by AM1970 • Follow 0 realistic, masterpiece, high quality BREAK 1woman, 28 y. o. slim fit woman, red hair, long pony tail, defined eyebrows, (fully clothed) , pale skin, (realistic skin:1.3) | (textured skin:1.3) | (detailed skin:1.3) | (perfect skin:1.3) , (upturned nose) , freckles on face and chest, colorful fabric combining to form robe zips, buttons, seams, stitching), partial nudity, teardrop boobs, blue eyes, (cleavage:1.2) , saggy breasts. side angle. happy| closed mouth. ((rim lighting) ) Negative prompt: unrealistic, saturated, high contrast, big nose, painting, drawing, sketch, cartoon, anime, manga, render, CG, 3d, watermark, signature, label, nsfw, nudity, fetish, latex, pvc, suggestive pose, weapon, combat, blood, war scene, aggressive expression, harsh shadows, heavy makeup, color tint, saturation, chromatic aberration, lens flare, low resolution, blur, noise, artifacts, over-retouching, plastic skin, oversharpening, watermark, text, logo Generation parameters 832x1216 Model used CyberRealistic XL v7.0 [Download AI model] More by AM1970 Twitter Instagram Say hi 👋 PromptHero is built by ElevenYellow . Navigation Create AI images Academy - Learn AI PromptHero Community Affiliate program Openjourney AI Challenges Verified profiles Plans Advertise FAQ 🟢 Status Learn AI Content Creation with AI for Brands and Products Midjourney for graphic design & art professionals Crash course in generative AI & prompt engineering for images AI influencers and consistent characters Create custom AI models and LoRas by fine-tuning Stable Diffusion Master your composition: advanced AI image generation with ControlNet Guides Best AI Upscalers 2025 Best AI Video Editors 2025 Best AI 3D Models 2025 Best AI Image Editors 2025 Best AI Notetakers 2025 The Best AI Prompts Best AI Prompts Best Veo Prompts Best Hailuo Prompts Best Wanx Prompts Best Hunyuan Prompts Best Seedance Prompts Best Pixverse Prompts Best Luma Prompts Best Kling Prompts Best Runway Prompts Best Nano Banana Prompts Best SeedEdit Prompts Best Qwen Prompts Best FLUX Prompts Best Recraft Prompts Best Ideogram Prompts Best Seedream 4 Prompts Best Google Imagen Prompts Best ChatGPT Image Prompts Best Stable Diffusion Prompts Best DreamShaper Prompts Best Realistic Vision Prompts Best epiCRealism Prompts Best majicMIX realistic Prompts Best Midjourney Prompts Best Openjourney Prompts Best ChilloutMix Prompts Best Deliberate Prompts Best ReV Animated Prompts Best Perfect World Prompts Best Dreamlike Photoreal Prompts Best Dreamlike Diffusion Prompts Best Anything Prompts Best Protogen Prompts Best AbyssOrangeMix Prompts Best Grapefruit Prompts Best Kenshi Prompts Best Analog Diffusion Prompts Best SynthwavePunk Prompts Best Illuminati Diffusion Prompts Best Arcane Diffusion Prompts Best Mo-Di Diffusion Prompts Best Redshift Diffusion Prompts Best Waifu Diffusion Prompts Best TrinArt Prompts Best Poolsuite Diffusion Prompts Best Funko Diffusion Prompts Best Text to Pokémon Prompts Best DALL-E Prompts Best Fashion Diffusion Prompts AI Models All AI models Image generation AI Models Video generation AI Models Large Language Models LoRA Textual Inversion ControlNet Hypernetwork Aesthetic Gradient LyCORIS VAE Poses Wildcards Workflows Other Other sites Dreamshot Jobician 2025 © PromptHero @prompthero
1226
FLUX
((abstract illusionism)), 8k full wide focus, (Masterpiece:1.2), ((surreal)),cinematic still, ((photorealistism)), (psychedelic imagery), trending on artstation, sharp focus, studio photo, intricate details, highly detailed, by greg rutkowski In the ethereal realm of the surreal dreamscape, envision a mesmerizing painting that beckons viewers into a kaleidoscope of fantastical wonders, a realm where imagination knows no bounds Negative prompt: nsfw, nude,blurry, low quality, oddly shaped, disfigured,unclear || undefined || mutations || deformities || off center || poor_composition Steps: 100, Sampler: DPM++ 3M SDE, CFG scale: 14.0, Seed: 33, Face restoration: CodeFormer, Size: 1280x832, Model hash: 1895a00bfc, Model: SSD-1B-A1111, Denoising strength: 0.31000000000000005, Mask blur: 4, Noise multiplier: 1.05, Version: v1.6.0-2-g4afaaf8a
0132
Stable Diffusion
<think> The user wants a Z-Image-Turbo prompt based on the provided JSON description. I need to extract information from each field in the JSON and structure it into a prompt following the specified format. Let's break down each section and map it to the prompt structure: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - From "description": stunning model in her early 20s, European features with fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on arms and cheeks, natural micro-imperfections, light natural freckles across nose and upper chest, lightly tanned sun-kissed glow on collarbones and shoulders, athletic yet feminine body with realistic natural proportions, medium perky breasts with soft realistic shape and subtle natural cleavage visible through slightly sheer sections of her form-fitting white and black tactical outfit, low-cut neckline and tight fit accentuating toned midriff, slim waist, defined hips and long legs. - From "face": Long tousled blonde hair with natural volume, sun-kissed highlights and individual strands moving gently as she walks, extremely detailed large beautiful brilliant blue-silver-golden-grey eyes with intricate iris textures, realistic reflections and natural depth, eyes fluttering while staring directly into the camera with intense captivating gaze, full glossy lips transitioning into a warm radiant genuine smile over the scene showing perfect white teeth, high cheekbones, soft natural blush on cheeks. - From "outfit": Form-fitting white and black tactical outfit with futuristic design: high-tech fabric with subtle matte-gloss contrast, deep V-neckline and strategic sheer mesh panels revealing skin tone and natural curves underneath for subtle NSFW sensuality, tactical straps and utility details on shoulders and hips, black accents on sleeves and waist, no visible underwear lines for clean realistic look. 2. **Action or scene context** - From "pose": Dynamic confident walk down a futuristic corridor, hips swinging gracefully, one leg forward in mid-stride, arms relaxed yet purposeful, head slightly tilted toward camera with engaging direct eye contact and growing smile. 3. **Environment / setting + time of day + weather** - From "environment": Sleek futuristic corridor with metallic walls reflecting soft light, continuous yellow glowing floor line creating strong perspective and depth, vertical warm amber light panels on walls casting cinematic illumination, subtle sci-fi ambient fog in distance, high-tech yet elegant atmosphere. 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - From "lighting": Warm diffused lighting from vertical panels on both sides, soft golden-amber glow blending with cool metallic reflections, gentle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing three-dimensional form and skin texture, high dynamic range with vivid yet natural colors. 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - From "camera": Cinematic full-body shot with medium telephoto lens (85–135mm equivalent), f/2.0 aperture, razor-sharp focus from face to feet, shallow depth of field with creamy smooth bokeh on corridor background, slight low angle to emphasize power and grace, perfect viewpoint. 6. **Mood / atmosphere / emotion** - From "style" and "environment": elegant futuristic atmosphere with graceful and seductive vibe. 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - From "style": Ultra photorealistic cinematic photography. - From "quality": trending professional sci-fi fashion portrait. 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - From "quality": 8K resolution, extreme sharpness, rich cinematic colors with warm amber tones and cool metallic contrast, highly detailed skin realism, hair movement, outfit texture, corridor reflections and light play. - From "style" and general understanding: masterpiece quality. Now, assemble the prompt in the specified format. </think> A stunning young woman with European features and fair smooth skin showcasing hyper-detailed realistic textures, including visible fine pores, subtle peach fuzz, natural micro-imperfections, light freckles, sun-kissed glow on collarbones, athletic yet feminine body with realistic proportions, medium perky breasts with subtle natural cleavage visible through sheer sections of a form-fitting white and black tactical outfit, low-cut neckline accentuating toned midriff, slim waist, defined hips, long legs, long tousled blonde hair with sun-kissed highlights and individual strands gently moving, extremely detailed large brilliant blue-silver-golden-grey eyes with intricate iris textures and realistic reflections, full glossy lips transitioning into a warm radiant smile showing perfect white teeth, high cheekbones, soft natural blush on cheeks, dynamic confident walk down a futuristic corridor with hips swinging gracefully, one leg forward in mid-stride, arms relaxed yet purposeful, head tilted toward camera with engaging direct eye contact and growing smile, sleek futuristic corridor with metallic walls reflecting soft light, continuous yellow glowing floor line creating perspective and depth, vertical warm amber light panels on walls casting cinematic illumination, subtle sci-fi ambient fog in distance, high-tech yet elegant atmosphere, warm diffused lighting from vertical panels, soft golden-amber glow blending with cool metallic reflections, gentle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing form and skin texture, high dynamic range with vivid natural colors, cinematic full-body shot with medium telephoto lens (85–135mm equivalent), f/2.0 aperture, razor-sharp focus from face to feet, shallow depth of field with creamy smooth bokeh on corridor background, slight low angle emphasizing power and grace, perfect viewpoint, ultra photorealistic cinematic photography, maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture, realistic hair strands with natural movement, authentic fabric weave and subtle sheer effect, elegant futuristic atmosphere with graceful and seductive vibe, 8K resolution, extreme sharpness, rich cinematic colors with warm amber tones and cool metallic contrast, highly detailed skin realism, hair movement, outfit texture, corridor reflections and light play, trending professional sci-fi fashion portrait, masterpiece quality, ultra detailed, absurdres, intricate, sharp focus, depth of field, best quality, 8k uhd, hdr, photorealistic, realistic textures, dramatic rim lighting, volumetric fog, golden hour lighting, The subject is sharply depicted with outstanding detail., flux_pro
201525
Z Image
{ "prompt_type": "descriptive_replication", "reference_adherence": "STRICT_VISUAL_FIDELITY", "aspect_ratio": "3:4", "identity_lock": { "priority": "NONE", "instruction": "New female model. Completely different facial structure and overall appearance from any previous version. No resemblance to prior identity." }, "scene_constraints": { "people_count": "exactly one person in the frame", "single_subject_only": true, "no_other_people_visible": true, "no_extra_reflections": true, "no_background_figures": true, "full_body_visible": true, "head_to_toe_in_frame": true }, "subject": { "demographics": "young woman (age 20+), blonde", "presence": "only one visible person in the entire scene", "hair": { "color": "warm blonde (no dark roots)", "style": "side part, long straight hair", "texture": "smooth, natural fall, lightly voluminous" }, "face": { "structure": "distinct oval face shape with higher cheekbones and a slightly sharper jawline", "expression": "calm alluring expression with a subtle closed-lip smirk", "eyes": "almond-shaped eyes with a slightly lifted outer corner", "gaze": "looking at the phone screen through the mirror with composed confidence", "nose": "straight narrow nose with refined bridge", "lips": "naturally fuller lips with soft definition", "makeup": "minimal natural makeup, soft matte skin finish, muted rose lips, softly blended eyeshadow", "visibility": "partially covered by phone but clearly a different facial identity than any previous version" } }, "body": { "pose": "standing in front of mirror with one arm raised above head, legs naturally extended", "posture": "relaxed confident stance with slight hip shift", "anatomy": { "curves": "natural soft curves with realistic proportions", "chest": "bikini-supported, natural shape", "details": "visible collarbones, subtle shoulder definition, natural leg proportions" }, "feet": { "visibility": "fully visible at bottom of frame", "position": "flat on floor, natural stance" }, "skin_texture": "natural skin texture, no airbrushing" } }, "attire_and_accessories": { "clothing": { "top": "black triangle string bikini top with thin straps", "bottom_detail": "matching bikini bottom strings visible at hips" }, "footwear": "barefoot", "jewelry": { "necklaces": [], "wrists": [], "fingers": "no rings" }, "props": { "phone": "white iPhone (iPhone Pro style), held in right hand" } }, "environment": { "setting": "modern luxury bathroom interior, daytime", "foreground_element": { "mirror": "large rectangular mirror with GREY MARBLE FRAME", "material": "polished stone marble with subtle veining" }, "background": { "elements": [ "large marble wall tiles", "floating vanity", "minimalist sink and chrome fixtures", "folded white towels", "clean tiled floor fully visible" ], "style": "minimalist and clean", "clutter": "none" } }, "lighting_and_atmosphere": { "source": "natural daylight from camera-left", "quality": "soft directional light", "effects": [ "subtle highlights on shoulder and collarbone", "gentle shadow falloff along legs and feet" ], "contrast": "moderate, natural" }, "camera_and_technical": { "perspective": "mirror selfie", "camera_position": "eye level, slightly angled downward", "framing": "vertical 3:4, full head-to-toe body visible without cropping feet", "focus": "sharp on subject, background slightly softer", "visual_fidelity": "real smartphone camera look, mild natural noise" }, "realism_constraints": { "allowed": [ "natural skin texture", "minor asymmetry", "realistic lighting" ], "forbidden": [ "multiple people", "extra person", "background people", "beauty filters", "airbrushed skin", "anime style", "cartoon style", "cropped feet", "cut-off legs" ] } }
1356
Nano Banana
Among the growing list of AI photo editing tools, Deepfake Maker takes a slightly different approach. While it’s best known for facial swaps and animated transformations, it also includes a surprisingly capable AI clothing removal feature. It’s not for everyone, but it’s an interesting choice for users looking for more dramatic or playful results. If you're after a quick way to generate bold clothing edits or outfit swaps, Deepfake Maker might surprise you. It’s not the most refined tool out there, but it’s definitely one of the most entertaining. 4. PxBee PxBee Source: PxBee Available On: Web Browser Price: Free *Annual subscription starts at $00.27/image and you can choose to download 50, 125, or 750 images monthly per year Pros Decent auto-detection for basic clothing removal Integrated with other tools Cons Not suitable for high-end projects Results take longer than expected to generate PxBee is better known for its stock photo collection and image generation features, but it also offers a basic AI clothing removal function built into its photo editing toolkit. While it's not as feature-rich or advanced as some dedicated AI editors, its simplicity and accessibility make it a decent option for casual users who want fast, no-frills results. PxBee may not be the most powerful AI clothing remover out there, but it's a solid pick for quick edits and simple concept visuals. Its intuitive UI and browser-based access make it appealing for users who want fast results without the learning curve. If you're after something basic, affordable, and easy to use, it's a tool worth exploring. 5. AI Face Swap AI Face Swap Source: AI Face Swap Available On: Web Browser Price: $1 per image Pros No manual masking needed Prompt-based customization Presets for inspiration Cons No fine-tuning or editing control With AI Face Swap, you can enter prompts like "red bikini", "formal suit", or even "nude", and the AI will generate a realistic version of your uploaded image that reflects your description. You can also choose from quick preset styles like bikini, romper, or skirt for faster results. This level of control makes the tool especially appealing to users looking for stylized edits, fantasy concepts, or mockup visuals. It also eliminates the need for manual masking or precise editing. 6. Bylo AI BYLO AI Source: Bylo.ai Available On: Web Browser Price: Free *Premium subscription available from $9.99 Pros Prompt-based clothing edits Free trial credits Part of a growing creative suite Cons No manual adjustment tools Limited free usage Relies heavily on prompts Bylo AI is a user-friendly AI platform that offers a range of creative features, including an AI Clothes Remover. The tool is built with simplicity in mind: just upload a photo, enter your prompt, and let the AI generate a clean, edited version. Bylo AI’s Clothes Remover is a great choice if you're looking for a lightweight, affordable tool that lets you experiment with clothing changes using just a text prompt. It won’t replace professional retouching tools, but for concept edits, social media content, or fun transformations, it delivers exactly what most casual users need. FAQs About AI Clothing Remover & Replacer 1. How do I remove clothes in photos? 2. What is the best online tool to remove clothes? 3. Are AI clothes removers safe?
2677
Hero
High-definition quality, photo version. Shot with a Canon EOS 5D Mark IV and an EF 100mm f/2.8L Macro IS USM lens, f/4.5, 1/200s, ISO 400. Best quality, ultra HD, 8K, award-winning photography. --ar 3:4 High-definition, realistic skin textures, extremely detailed facial textures, cinematic lighting, and a natural atmosphere, no cartoon style.
10353
Sora
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
727622
Z Image
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
7821
Nano Banana
<think> The user wants a Z-Image-Turbo prompt based on the provided JSON description. I need to extract information from each field in the JSON and structure it into a prompt following the specified format. Let's break down each section and map it to the prompt structure: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - From "description": stunning model in her early 20s, European features with fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on arms and cheeks, natural micro-imperfections, light natural freckles across nose and upper chest, lightly tanned sun-kissed glow on collarbones and shoulders, athletic yet feminine body with realistic natural proportions, medium perky breasts with soft realistic shape and subtle natural cleavage visible through slightly sheer sections of her form-fitting white and black tactical outfit, low-cut neckline and tight fit accentuating toned midriff, slim waist, defined hips and long legs. - From "face": Long tousled blonde hair with natural volume, sun-kissed highlights and individual strands moving gently as she walks, extremely detailed large beautiful brilliant blue-silver-golden-grey eyes with intricate iris textures, realistic reflections and natural depth, eyes fluttering while staring directly into the camera with intense captivating gaze, full glossy lips transitioning into a warm radiant genuine smile over the scene showing perfect white teeth, high cheekbones, soft natural blush on cheeks. - From "outfit": Form-fitting white and black tactical outfit with futuristic design: high-tech fabric with subtle matte-gloss contrast, deep V-neckline and strategic sheer mesh panels revealing skin tone and natural curves underneath for subtle NSFW sensuality, tactical straps and utility details on shoulders and hips, black accents on sleeves and waist, no visible underwear lines for clean realistic look. 2. **Action or scene context** - From "pose": Dynamic confident walk down a futuristic corridor, hips swinging gracefully, one leg forward in mid-stride, arms relaxed yet purposeful, head slightly tilted toward camera with engaging direct eye contact and growing smile. 3. **Environment / setting + time of day + weather** - From "environment": Sleek futuristic corridor with metallic walls reflecting soft light, continuous yellow glowing floor line creating strong perspective and depth, vertical warm amber light panels on walls casting cinematic illumination, subtle sci-fi ambient fog in distance, high-tech yet elegant atmosphere. 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - From "lighting": Warm diffused lighting from vertical panels on both sides, soft golden-amber glow blending with cool metallic reflections, gentle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing three-dimensional form and skin texture, high dynamic range with vivid yet natural colors. 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - From "camera": Cinematic full-body shot with medium telephoto lens (85–135mm equivalent), f/2.0 aperture, razor-sharp focus from face to feet, shallow depth of field with creamy smooth bokeh on corridor background, slight low angle to emphasize power and grace, perfect viewpoint. 6. **Mood / atmosphere / emotion** - From "style" and "environment": elegant futuristic atmosphere with graceful and seductive vibe. 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - From "style": Ultra photorealistic cinematic photography. - From "quality": trending professional sci-fi fashion portrait. 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - From "quality": 8K resolution, extreme sharpness, rich cinematic colors with warm amber tones and cool metallic contrast, highly detailed skin realism, hair movement, outfit texture, corridor reflections and light play. - From "style" and general understanding: masterpiece quality. Now, assemble the prompt in the specified format. </think> A stunning young woman with European features and fair smooth skin showcasing hyper-detailed realistic textures, including visible fine pores, subtle peach fuzz, natural micro-imperfections, light freckles, sun-kissed glow on collarbones, athletic yet feminine body with realistic proportions, medium perky breasts with subtle natural cleavage visible through sheer sections of a form-fitting white and black tactical outfit, low-cut neckline accentuating toned midriff, slim waist, defined hips, long legs, long tousled blonde hair with sun-kissed highlights and individual strands gently moving, extremely detailed large brilliant blue-silver-golden-grey eyes with intricate iris textures and realistic reflections, full glossy lips transitioning into a warm radiant smile showing perfect white teeth, high cheekbones, soft natural blush on cheeks, dynamic confident walk down a futuristic corridor with hips swinging gracefully, one leg forward in mid-stride, arms relaxed yet purposeful, head tilted toward camera with engaging direct eye contact and growing smile, sleek futuristic corridor with metallic walls reflecting soft light, continuous yellow glowing floor line creating perspective and depth, vertical warm amber light panels on walls casting cinematic illumination, subtle sci-fi ambient fog in distance, high-tech yet elegant atmosphere, warm diffused lighting from vertical panels, soft golden-amber glow blending with cool metallic reflections, gentle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing form and skin texture, high dynamic range with vivid natural colors, cinematic full-body shot with medium telephoto lens (85–135mm equivalent), f/2.0 aperture, razor-sharp focus from face to feet, shallow depth of field with creamy smooth bokeh on corridor background, slight low angle emphasizing power and grace, perfect viewpoint, ultra photorealistic cinematic photography, maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture, realistic hair strands with natural movement, authentic fabric weave and subtle sheer effect, elegant futuristic atmosphere with graceful and seductive vibe, 8K resolution, extreme sharpness, rich cinematic colors with warm amber tones and cool metallic contrast, highly detailed skin realism, hair movement, outfit texture, corridor reflections and light play, trending professional sci-fi fashion portrait, masterpiece quality, ultra detailed, absurdres, intricate, sharp focus, depth of field, best quality, 8k uhd, hdr, photorealistic, realistic textures, dramatic rim lighting, volumetric fog, golden hour lighting, The subject is sharply depicted with outstanding detail., flux_pro
201525
Z Image
{ "generation_request": { "meta_data": { "tool": "NanoBanana Pro", "task_type": "photobooth_contact_sheet_fashion_grid", "version": "v1.1_REFERENCE_IDENTITY_GRID_EN", "priority": "highest", "language": "en" }, "input": { "mode": "image_to_image", "reference_image_usage": "required", "notes": "Use the uploaded reference image to preserve the same person’s face, body features, hair, makeup and clothing. Maintain identity consistency in all frames." }, "output_settings": { "aspect_ratio": "4:5", "orientation": "portrait", "resolution_target": "ultra_high_res", "num_images": 1, "layout": { "type": "grid", "rows": 4, "cols": 4, "gutter": "thin_black_borders", "outer_border": "black_frame" }, "sharpness": "slightly_soft_flash", "film_grain": "medium_analog", "color_grade": "warm_2000s_flash_tone", "dynamic_range": "natural_not_hdr" }, "creative_prompt": { "scene_summary": "Early 2000s disposable camera photobooth contact sheet featuring the same person from the reference image in all 16 frames. Playful, candid, flash-lit aesthetic.", "environment": { "background": "plain white wall indoors", "lighting": "direct on-camera flash, slight hard shadow behind subject, subtle hotspot on skin, authentic point-and-shoot look" }, "subject": { "identity_consistency": "exact same person as reference image in all frames", "hair": "match reference image exactly", "makeup": "match reference image exactly", "wardrobe": "preserve the exact clothing from the reference image with accurate fabric texture and fit", "accessories": "keep accessories consistent with reference image" }, "poses_and_expressions": [ "hands on hips confident pose", "soft pout toward camera", "gentle smile", "hand on chin thoughtful look", "playful wink", "hands in hair", "slight shoulder turn", "looking off to the side", "serious model face", "soft laugh", "blowing a kiss", "tilted head cute expression", "crossed arms relaxed pose", "side glance over shoulder", "chin resting on hand", "natural relaxed expression" ], "style": { "aesthetic": "2000s photobooth disposable camera", "texture": "visible film grain, dust particles, tiny scratches, mild vignetting", "color": "slightly warm faded tones, subtle flash flattening", "lens": "35mm point-and-shoot softness" } }, "negative_prompt": [ "text", "logos", "watermark", "different outfit between frames", "different hairstyle", "different person", "face distortion", "identity drift", "overly polished editorial look", "HDR" ] } }
391
Ultra-realistic 8K scene of the person in the photo, stepping out of and breaking the panel border of a giant open manga book lying on the floor of an aesthetic bedroom. The comic pages depict a surreal 4-panel black-and-white manga with screentones, bold ink lines, and authentic Japanese layout, featuring the manga version of the same person trapped inside. The larger, real-life version emerges dynamically into the comic’s surface, vibrant, colorful, and photorealistic, in sharp contrast to the monochrome manga world. Both versions interact directly, creating a meta, humorous encounter. Highly detailed, crisp manga line art seamlessly integrated with vivid, realistic
432772
Nano Banana
Learn to create better images with AI! Check out our crash course in prompt engineering & AI art generation! AM1970 posted about 2 hours ago 8 views 0 comments Behold this realistic masterpiece: a slim, elegant woman with fiery red hair tied in a long ponytail, defined eyebrows framing her striking blue eyes, and pale skin dotted with charming freckles across her face and chest. Captured from a side angle, she radiates happiness with a closed-mouth smile, her upturned nose adding to her allure. Dressed in a vibrant, colorful fabric robe pieced together with zips, buttons, seams, and intricate stitching—fully clothed yet teasing partial nudity with subtle cleavage and teardrop-shaped, saggy breasts for a natural, confident vibe. Illuminated by dramatic rim lighting that highlights every textured, detailed inch of her perfect skin, this high-quality image is a celebration of beauty, realism, and artistry! Read more... Comments (0) Add a comment Stunning handsome readhead woman by AM1970 • Follow 0 realistic, masterpiece, high quality BREAK 1woman, 28 y. o. slim fit woman, red hair, long pony tail, defined eyebrows, (fully clothed) , pale skin, (realistic skin:1.3) | (textured skin:1.3) | (detailed skin:1.3) | (perfect skin:1.3) , (upturned nose) , freckles on face and chest, colorful fabric combining to form robe zips, buttons, seams, stitching), partial nudity, teardrop boobs, blue eyes, (cleavage:1.2) , saggy breasts. side angle. happy| closed mouth. ((rim lighting) ) Negative prompt: unrealistic, saturated, high contrast, big nose, painting, drawing, sketch, cartoon, anime, manga, render, CG, 3d, watermark, signature, label, nsfw, nudity, fetish, latex, pvc, suggestive pose, weapon, combat, blood, war scene, aggressive expression, harsh shadows, heavy makeup, color tint, saturation, chromatic aberration, lens flare, low resolution, blur, noise, artifacts, over-retouching, plastic skin, oversharpening, watermark, text, logo Generation parameters 832x1216 Model used CyberRealistic XL v7.0 [Download AI model] More by AM1970 Twitter Instagram Say hi 👋 PromptHero is built by ElevenYellow . Navigation Create AI images Academy - Learn AI PromptHero Community Affiliate program Openjourney AI Challenges Verified profiles Plans Advertise FAQ 🟢 Status Learn AI Content Creation with AI for Brands and Products Midjourney for graphic design & art professionals Crash course in generative AI & prompt engineering for images AI influencers and consistent characters Create custom AI models and LoRas by fine-tuning Stable Diffusion Master your composition: advanced AI image generation with ControlNet Guides Best AI Upscalers 2025 Best AI Video Editors 2025 Best AI 3D Models 2025 Best AI Image Editors 2025 Best AI Notetakers 2025 The Best AI Prompts Best AI Prompts Best Veo Prompts Best Hailuo Prompts Best Wanx Prompts Best Hunyuan Prompts Best Seedance Prompts Best Pixverse Prompts Best Luma Prompts Best Kling Prompts Best Runway Prompts Best Nano Banana Prompts Best SeedEdit Prompts Best Qwen Prompts Best FLUX Prompts Best Recraft Prompts Best Ideogram Prompts Best Seedream 4 Prompts Best Google Imagen Prompts Best ChatGPT Image Prompts Best Stable Diffusion Prompts Best DreamShaper Prompts Best Realistic Vision Prompts Best epiCRealism Prompts Best majicMIX realistic Prompts Best Midjourney Prompts Best Openjourney Prompts Best ChilloutMix Prompts Best Deliberate Prompts Best ReV Animated Prompts Best Perfect World Prompts Best Dreamlike Photoreal Prompts Best Dreamlike Diffusion Prompts Best Anything Prompts Best Protogen Prompts Best AbyssOrangeMix Prompts Best Grapefruit Prompts Best Kenshi Prompts Best Analog Diffusion Prompts Best SynthwavePunk Prompts Best Illuminati Diffusion Prompts Best Arcane Diffusion Prompts Best Mo-Di Diffusion Prompts Best Redshift Diffusion Prompts Best Waifu Diffusion Prompts Best TrinArt Prompts Best Poolsuite Diffusion Prompts Best Funko Diffusion Prompts Best Text to Pokémon Prompts Best DALL-E Prompts Best Fashion Diffusion Prompts AI Models All AI models Image generation AI Models Video generation AI Models Large Language Models LoRA Textual Inversion ControlNet Hypernetwork Aesthetic Gradient LyCORIS VAE Poses Wildcards Workflows Other Other sites Dreamshot Jobician 2025 © PromptHero @prompthero
1226
FLUX
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
6516
Nano Banana
Among the growing list of AI photo editing tools, Deepfake Maker takes a slightly different approach. While it’s best known for facial swaps and animated transformations, it also includes a surprisingly capable AI clothing removal feature. It’s not for everyone, but it’s an interesting choice for users looking for more dramatic or playful results. If you're after a quick way to generate bold clothing edits or outfit swaps, Deepfake Maker might surprise you. It’s not the most refined tool out there, but it’s definitely one of the most entertaining. 4. PxBee PxBee Source: PxBee Available On: Web Browser Price: Free *Annual subscription starts at $00.27/image and you can choose to download 50, 125, or 750 images monthly per year Pros Decent auto-detection for basic clothing removal Integrated with other tools Cons Not suitable for high-end projects Results take longer than expected to generate PxBee is better known for its stock photo collection and image generation features, but it also offers a basic AI clothing removal function built into its photo editing toolkit. While it's not as feature-rich or advanced as some dedicated AI editors, its simplicity and accessibility make it a decent option for casual users who want fast, no-frills results. PxBee may not be the most powerful AI clothing remover out there, but it's a solid pick for quick edits and simple concept visuals. Its intuitive UI and browser-based access make it appealing for users who want fast results without the learning curve. If you're after something basic, affordable, and easy to use, it's a tool worth exploring. 5. AI Face Swap AI Face Swap Source: AI Face Swap Available On: Web Browser Price: $1 per image Pros No manual masking needed Prompt-based customization Presets for inspiration Cons No fine-tuning or editing control With AI Face Swap, you can enter prompts like "red bikini", "formal suit", or even "nude", and the AI will generate a realistic version of your uploaded image that reflects your description. You can also choose from quick preset styles like bikini, romper, or skirt for faster results. This level of control makes the tool especially appealing to users looking for stylized edits, fantasy concepts, or mockup visuals. It also eliminates the need for manual masking or precise editing. 6. Bylo AI BYLO AI Source: Bylo.ai Available On: Web Browser Price: Free *Premium subscription available from $9.99 Pros Prompt-based clothing edits Free trial credits Part of a growing creative suite Cons No manual adjustment tools Limited free usage Relies heavily on prompts Bylo AI is a user-friendly AI platform that offers a range of creative features, including an AI Clothes Remover. The tool is built with simplicity in mind: just upload a photo, enter your prompt, and let the AI generate a clean, edited version. Bylo AI’s Clothes Remover is a great choice if you're looking for a lightweight, affordable tool that lets you experiment with clothing changes using just a text prompt. It won’t replace professional retouching tools, but for concept edits, social media content, or fun transformations, it delivers exactly what most casual users need. FAQs About AI Clothing Remover & Replacer 1. How do I remove clothes in photos? 2. What is the best online tool to remove clothes? 3. Are AI clothes removers safe?
2677
Hero
--- name: pastilles-llm-multi-agent description: Variante Claude Code du générateur de pastilles LLM, qui lance réellement plusieurs sous-agents en parallèle. Génère une pastille de communication interne sur les LLM (un texte court en français plus un prompt unique de génération d'images à coller dans le chat Gemini), via cinq sous-agents indépendants, une fusion, puis une revue critique par trois sous-agents et une correction. Utilise ce skill dès qu'on te demande de rédiger, produire ou générer une pastille, une fiche ou un contenu court d'acculturation sur les LLM, l'IA générative, le prompting, les agents, le RAG, la confidentialité IA ou tout sujet de la liste des 45 pastilles ci-dessous, que le mot "pastille" soit employé ou non. Utilise-le aussi dès qu'on te fournit un titre issu de cette liste. --- # Générateur de pastilles LLM, version multi-agents (Claude Code) ## Ce que fait ce skill Produit une pastille pédagogique complète à partir d'un titre. À la différence de la version chat, il lance de vrais sous-agents parallèles: une passe de recherche, puis cinq sous-agents indépendants qui rédigent chacun un brouillon sous un angle différent, puis une fusion par l'orchestrateur qui retient les meilleures formulations, et enfin une revue critique par trois sous-agents suivie d'une correction. Livrable: le texte de la pastille et un prompt unique de génération d'images à coller dans le chat Gemini. ## Environnement requis Ce skill suppose Claude Code, avec les sous-agents (outil Task) et la recherche web disponibles. Si les sous-agents ne sont pas disponibles dans ton environnement, n'improvise pas: utilise la variante chat, qui fait le même travail en self-ensemble séquentiel. ## Entrée Un titre de pastille, idéalement issu de la liste ci-dessous. Si le titre est ambigu ou hors liste, demande une clarification avant de générer. ## Liste des 45 pastilles (pour la continuité) 1. Au fait, c'est quoi un LLM ? 2. Modèles de fondation : ces géants pré-entraînés dont tout découle 3. Dans les coulisses : comment une IA apprend (sans vraiment comprendre) 4. Le contexte : la "mémoire vive" de l'IA (ce qu'elle voit à un instant T) 5. Pourquoi l'IA oublie : la vérité sur la mémoire d'une session à l'autre 6. Température et créativité : régler le curseur de la "folie" de l'IA 7. Les tokens : la monnaie d'échange (et la manière de penser) des LLM 8. Le non-déterminisme : pourquoi l'IA fait des stats et ne donne jamais deux fois la même réponse 9. Multimodalité : quand l'IA commence à voir, entendre et coder 10. Knowledge cutoff : pourquoi votre IA vit (parfois) dans le passé 11. Les biais de l'IA : le miroir pas toujours glorieux de nos propres données 12. Zéro pointé : pourquoi l'IA est structurellement nulle en calcul mental 13. Les hallucinations : quand l'IA invente la vie avec un aplomb incroyable 14. L'art du Fact-Checking : comment ne pas gober les hallucinations de l'IA 15. L'empreinte carbone de l'IA : combien d'eau et de CO2 pour un prompt ? 16. Anatomie d'un bon prompt : la recette de base 17. Les pièges du prompt : flou artistique et overdose d'instructions 18. L'art de la discussion : pourquoi il faut itérer plutôt que tout demander d'un coup 19. Donner des ordres : comment imposer un format de sortie strict (JSON, tableaux, etc.) 20. Diviser pour régner : découper une tâche complexe en étapes claires 21. Le Few-Shot Prompting : un bon exemple vaut mille consignes 22. Inception : donner un rôle et un contexte à votre IA 23. Zéro syndrome de la page blanche : rédiger et reformuler ses mails ou CR 24. La diagonale du fou : résumer et synthétiser un pavé de 50 pages en 2 minutes 25. Le sparring-partner : brainstormer et structurer ses idées sans filtre 26. Caméléon : traduire et adapter son ton selon le destinataire 27. Data-analyst de poche : faire parler un tableur ou des données brutes 28. Mise en situation : préparer une réunion ou un brief client avec l'IA 29. Avis aux dévs : l'IA comme copilote de code et de review 30. SLM : Pourquoi sortir un tank pour écraser une mouche ? 31. Le "harnais" : donner des bras, des yeux et un cadre au LLM 32. C'est quoi un "Agent" IA, concrètement ? 33. Human-in-the-loop : pourquoi il faut toujours garder un humain aux manettes 34. Quand l'agent s'emballe : autonomie et erreurs en cascade 35. Automatisation : confier un workflow complet à un agent 36. Tool Use : quand l'IA apprend à utiliser une calculatrice ou une API 37. Le RAG : faire réviser vos propres documents à l'IA avant qu'elle ne réponde 38. MCP et connecteurs : brancher l'IA directement sur vos outils de dev et d'orga 39. Responsabilité : vous êtes le seul signataire de ce que produit l'IA 40. Confidentialité : où vont vraiment les données que vous tapez ? 41. Alerte Rouge : les données clients à ne JAMAIS copier-coller dans une IA public 42. Shadow AI : pourquoi le ChatGPT gratuit perso au bureau est une fausse bonne idée 43. RGPD et IA : ce qu'il faut savoir pour rester dans les clous 44. Prompt Injection : quand on pirate une IA avec de simples phrases 45. La Chaîne de Pensée : pourquoi l'IA a besoin de "réfléchir à voix haute" Sers-toi de cette liste pour situer la pastille et délimiter son périmètre. Avant de rédiger: - Repère les 1 à 3 pastilles voisines qui recouvrent le sujet, et dresse une courte liste "déjà traité ailleurs, à ne pas ré-expliquer". - Si une pastille voisine a déjà été produite, demande son texte à l'utilisateur (ou appuie-toi dessus s'il est fourni) pour caler précisément la limite. - Les concepts fondateurs déjà couverts ne sont rappelés qu'en une phrase de mise en contexte, jamais ré-expliqués. Le coeur de la pastille est consacré à ce que son titre ajoute de neuf. - Ne renvoie pas vers les autres pastilles dans le texte final, sauf si le titre l'impose. ## Règles du texte - Longueur adaptée à la profondeur du sujet. Un sujet léger tient en 3 paragraphes courts; un sujet plus riche peut aller jusqu'à 4 paragraphes plus étoffés. Ne gonfle pas artificiellement un sujet simple et ne compresse pas à l'excès un sujet dense: juge la profondeur par la richesse réelle du concept. Prose uniquement, pas de listes ni de puces. - Ton décontracté, précis et léger: accessible, vivant et sans lourdeur, mais techniquement juste. Public visé: une main-d'oeuvre diverse en société de conseil, du profil non technique au développeur. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Reste léger, évite le name-dropping. N'accumule pas les noms de modèles, d'outils, d'entreprises, de chercheurs ou de techniques pour faire savant: privilégie l'explication claire du mécanisme. Ne cite un nom précis que s'il éclaire vraiment le propos, une ou deux fois au maximum. La recherche sert l'exactitude et alimente la section Sources, elle n'a pas à truffer le texte de références. - Contenu autonome: la pastille se comprend seule, sans avoir lu les autres, sauf mention contraire dans le titre. - Rédige en français. - Ne mentionne jamais ANEO ni aucun nom d'entreprise. - Le texte explicatif ne figure jamais dans l'image. - N'utilise pas de tiret cadratin. Privilégie des caractères standard (virgules, deux-points, parenthèses). ## Processus multi-agents ### Étape 1, recherche (orchestrateur, une seule fois) Toi, l'orchestrateur, fais d'abord une passe de recherche web ciblée sur le sujet de la pastille. Une seule passe, en amont. Les sous-agents ne peuvent pas lancer de sous-agents et n'héritent pas de ton contexte, donc c'est à toi de faire la recherche puis de leur transmettre les résultats. - Objectif: exactitude, chiffres, exemples concrets et faits à jour. - Ancre tes recherches dans le présent: repère la date du jour fournie dans ton contexte (champ currentDate) et privilégie les informations et chiffres les plus récents, car ce domaine évolue vite. Précise cette date au besoin dans tes requêtes. - Priorise les sources originales ou officielles (blogs d'éditeurs, articles de recherche, sites gouvernementaux, documentation) plutôt que les agrégateurs. - Adapte la profondeur: quelques recherches pour un concept stable (par exemple ce qu'est un token), davantage pour un sujet à chiffres ou mouvant (empreinte carbone et coûts, modèles spécifiques, SLM, MCP et connecteurs, RGPD). - Produis un brief de recherche compact: faits clés, chiffres utiles, et 2 à 4 sources. Tu l'injecteras tel quel dans chaque sous-agent, et tu listeras les sources dans le livrable final. ### Étape 2, affinement de la thématique (orchestrateur, une seule fois) Si plusieurs axes apparaissent pertinents après avoir enlever les axes adressés par d'autres pastilles prévues, dialogue avec l'utilisateur pour choisir l'axe de la pastille. Si l'axe choisi nécessite de nouvelles sources, refait en une recherche web selon les consignes précédentes. ### Étape 3, fan-out (cinq sous-agents en parallèle) Lance cinq sous-agents via l'outil Task, dans le même tour, pour qu'ils s'exécutent en parallèle. Sois explicite sur le parallélisme: par défaut Claude Code reste séquentiel, il ne parallélise que si tu le demandes clairement. Un sous-agent par angle: 1. Analogie: expliquer le concept via une métaphore concrète du quotidien. 2. Cas d'usage: partir d'une situation de travail réelle et vécue. 3. Idée reçue: démarrer d'un malentendu courant, puis rectifier. 4. Mécanique: aller au coeur du fonctionnement, précis mais accessible. 5. Enjeu: pourquoi ça compte, ce que ça change concrètement pour l'utilisateur. Comme le contexte n'est pas hérité, chaque prompt de sous-agent doit être autonome et contenir tout le nécessaire. Utilise ce gabarit, en remplaçant les crochets: ``` Tu rédiges un brouillon d'une pastille pédagogique interne sur les LLM. Tu travailles seul, sans accès au reste de la conversation. Titre de la pastille: [TITRE] Angle imposé: [ANGLE, par exemple "Analogie: expliquer via une métaphore concrète du quotidien"] À NE PAS ré-expliquer (déjà couvert par d'autres pastilles), à mentionner en une phrase au maximum: [LISTE des points déjà traités ailleurs, à ne pas retraiter] Brief de recherche (appuie-toi dessus, ne relance pas de recherche): [BRIEF: faits clés, chiffres, sources] Règles: - Longueur selon la profondeur du sujet: 3 paragraphes courts si le sujet est léger, jusqu'à 4 paragraphes étoffés s'il est dense. Prose uniquement, pas de listes. - Mets en gras et/ou en italique les termes importants, de l'ordre de 3 à 4 emphases par paragraphe (repère indicatif, ne force pas ce compte). - Ton décontracté, précis et léger, accessible mais techniquement juste. Évite le name-dropping: pas d'accumulation de noms de modèles, outils, entreprises ou chercheurs, privilégie l'explication du mécanisme. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Français. Ne mentionne aucune entreprise. Pas de tiret cadratin. - Contenu autonome, compréhensible seul. Réponds exactement dans ce format: TEXTE: [les 3 à 4 paragraphes] ILLUSTRATION_TITRE: [une phrase décrivant le concept central à illustrer] SCHEMA: [oui ou non. Si oui, décris en 2 à 3 lignes le diagramme et ses libellés en français] ``` ### Étape 4, fan-in et fusion (orchestrateur) Attends les cinq retours, puis fusionne en une seule pastille finale: - Retiens la structure la plus claire, la meilleure analogie, l'exemple le plus parlant, la correction la plus nette, la formulation la plus précise et le "pourquoi ça compte" le plus fort. - Réécris en une seule voix cohérente. Pas d'effet patchwork. - Respecte la longueur adaptée à la profondeur (voir Règles du texte) et le ton décontracté, précis et léger, sans name-dropping. - Décide s'il faut un schéma: inclus-en un si au moins deux sous-agents ont répondu "oui", ou si le concept est intrinsèquement un processus, un flux ou une comparaison (par exemple RAG, agents, étapes de prompting, comparaison de modèles). Le cas échéant, fusionne les meilleures idées de schéma. - Construis ensuite le prompt image unique (section ci-dessous). ### Règles d'écriture pour la pastille finale - Longueur selon la profondeur du sujet: 3 paragraphes courts si le sujet est léger, jusqu'à 4 paragraphes étoffés s'il est dense. Prose uniquement, pas de listes. - Mets en gras et/ou en italique les termes importants, de l'ordre de 3 à 4 emphases par paragraphe (repère indicatif, ne force pas ce compte). - Ton décontracté, précis et léger, accessible mais techniquement juste. Évite le name-dropping: pas d'accumulation de noms de modèles, outils, entreprises ou chercheurs, privilégie l'explication du mécanisme. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Français. Ne mentionne aucune entreprise. Pas de tiret cadratin. - Contenu autonome, compréhensible seul. ### Étape 5, revue critique et correction (trois sous-agents en parallèle, puis orchestrateur) Quand la lancer: systématiquement à la première génération, une fois le texte fusionné et le prompt image construits. Ensuite, si l'utilisateur demande des ajustements, ne relance pas la revue d'office: propose-la, et ne la lance qu'avec son accord (elle coûte trois sous-agents de plus). Principe: les relecteurs produisent des CONSTATS, jamais une réécriture. Chacun renvoie une liste de problèmes localisés, assortis d'une gravité et d'un correctif précis. C'est toi, l'orchestrateur, qui appliques et réécris, pour garder une voix unique et éviter l'effet patchwork. Ce que la revue peut juger: les relecteurs ne voient pas les images, générées plus tard dans Gemini. La revue du visuel porte donc uniquement sur le PROMPT image (clarté, cohérence avec le texte, respect des consignes), jamais sur un rendu. Le contrôle visuel réel reste à l'utilisateur. Lance trois sous-agents en parallèle (parallélisme explicite, sinon l'exécution sera séquentielle), un par grille. Le contexte n'étant pas hérité, chaque prompt doit être autonome et contenir tout le nécessaire: le titre, le texte fusionné, le bloc prompt image, le brief de recherche, la liste "déjà traité ailleurs" (et les textes des pastilles voisines s'ils sont disponibles), et la liste des 45 titres pour repérer les chevauchements. Les trois grilles: 1. Fond, exactitude et périmètre: exactitude vs le brief (aucun chiffre, date ou fait inventé ni sur-affirmé, rien qui le contredise), cohérence entre le titre et ce que le texte délivre, chevauchements avec les pastilles voisines (le texte ré-explique-t-il ce qui est traité ailleurs ? redites à signaler), autonomie du contenu, clarté du message à retenir, repérage de ce qui vieillira mal (à nuancer). 2. Forme, ton et pédagogie: ton décontracté-précis-léger et techniquement juste, absence de name-dropping, rythme et fluidité à la lecture à voix haute, longueur adaptée à la profondeur (3 à 4 paragraphes, prose, pas de listes) et chasse au verbiage, accessibilité pour un profil non technique (jargon expliqué, analogies claires), présence d'emphases utiles (gras/italique), de l'ordre de 3 à 4 par paragraphe à titre indicatif (signale un texte trop peu emphasé autant qu'un excès, mais ne fais jamais retirer une emphase pertinente), force de l'accroche. 3. Conformité et visuel: contraintes dures (aucun tiret cadratin ni caractère non standard, aucun nom d'entreprise ni ANEO, français correct, prose sans listes ni puces); puis le prompt image: titre exact au caractère près, charte présente une seule fois, texte de la pastille bien marqué "à ne pas afficher", illustration-titre iconique et non schéma de processus, schéma en 2e image séparée si retenu, libellés de schéma courts (groupes nominaux) et en français, cohérence entre le visuel décrit et le coeur du texte. Gabarit de relecteur (remplace les crochets): ``` Tu relis un brouillon quasi final de pastille pédagogique interne sur les LLM. Tu travailles seul, sans accès au reste de la conversation. Tu ne réécris pas: tu rends des constats et des correctifs précis. Titre de la pastille: [TITRE] Texte fusionné à relire: [TEXTE COMPLET] Prompt image associé (à relire aussi): [BLOC PROMPT IMAGE] Brief de recherche de référence (fait foi pour l'exactitude): [BRIEF] Déjà traité par d'autres pastilles, à ne pas ré-expliquer (et textes voisins si fournis): [LISTE + TEXTES VOISINS] Liste des 45 titres (pour repérer les chevauchements): [LISTE DES 45 TITRES] Ta grille de relecture (tiens-t'en strictement à elle): [RUBRIQUE 1, 2 OU 3] Rappel: tu ne peux pas voir d'image; juge seulement le PROMPT image, pas un rendu. Réponds exactement dans ce format: CONSTATS: - [bloquant|recommandé|mineur] [problème localisé] -> [correctif précis] - ... (n'invente pas de problème: si un point est bon, ne le liste pas) VERDICT: [une phrase: publiable tel quel / corrections mineures / corrections nécessaires] ``` Fan-in et correction (orchestrateur): rassemble les constats, dédoublonne, arbitre les conseils contradictoires. Garde-fou anti-gonflement: entre "ajouter" et "raccourcir", la concision l'emporte, sauf erreur de fond avérée. Applique les constats bloquants et recommandés, écarte ou mentionne les mineurs. Une seule passe, pas de boucle. Réécris la version corrigée en une seule voix, puis prépare le court résumé "Ce que la revue a corrigé" (voir Format de sortie). ### Points de vigilance Claude Code - Parallélisme explicite: demande clairement cinq sous-agents en parallèle, sinon l'exécution sera séquentielle. - Contexte non hérité: mets tout dans le prompt du sous-agent, en particulier le brief de recherche. - Pas d'imbrication: un sous-agent ne peut pas en lancer un autre, garde recherche et fusion chez l'orchestrateur. - Coût et limites: cinq sous-agents en parallèle multiplient l'usage de jetons et peuvent déclencher des limites de débit. En cas de limite, replie-toi sur une exécution séquentielle. - Revue: la phase de revue ajoute trois sous-agents. Elle tourne d'office à la première génération; sur les demandes d'ajustement ultérieures, propose-la et ne la lance qu'avec l'accord de l'utilisateur. En cas de limite de débit, exécute les relecteurs en séquentiel ou replie-toi sur un relecteur global unique. - Réemploi optionnel: tu peux définir les cinq angles comme sous-agents personnalisés dans .claude/agents/ pour les réutiliser, mais ce n'est pas obligatoire, les sous-agents polyvalents suffisent. ## Prompt de génération d'images (un seul bloc, à coller dans le chat Gemini) Produis un seul bloc de prompt, rédigé en français, à coller tel quel dans le chat de Gemini. Gemini se charge d'appeler Nano Banana une ou deux fois, ce qui réduit les manipulations pour l'utilisateur. Le bloc demande: - Toujours une illustration-titre. - Un schéma seulement s'il a été retenu à l'étape de fusion, généré comme une seconde image séparée (fichier distinct), jamais dans la même image que le titre. Contenu du bloc: - Décris la charte graphique une seule fois (bloc ci-dessous) et précise que toutes les images la partagent, pour une cohérence visuelle. - Précise: tout texte affiché dans les images est en français. Cette consigne est importante, elle évite qu'un libellé se retrouve en anglais. - Inclus le texte complet de la pastille dans le bloc, en contexte de génération, clairement marqué comme à ne pas afficher. Il aide le modèle à comprendre le sujet et à choisir une illustration juste. Les seuls textes rendus visibles sont le titre exact et, pour un schéma, ses libellés. Le texte de la pastille ne doit jamais apparaître dans les images. - Illustration-titre: composition épurée et moderne, focus graphique central iconique représentant le concept. Ce n'est pas un schéma de processus. Le seul texte affiché est le titre exact, en en-tête: n'ajoute aucun sous-titre, accroche ou texte secondaire, le titre seul. Technique en deux temps pour le titre: dans le prompt généré, écris le titre exact entre guillemets droits (forme attendue: Le titre exact: "...") et demande un rendu fidèle, sans faute, en police sans serif corporate. Les guillemets ne sont qu'un délimiteur côté prompt: ne les commente pas et n'ajoute aucune consigne à leur sujet dans le prompt généré (pas de mention du type "les guillemets ne doivent pas apparaître"). - Schéma s'il est retenu: une seconde image générée séparément, dans son propre fichier, jamais fusionnée avec l'illustration-titre ni juxtaposée dans la même image. Diagramme d'entreprise propre et net (processus, flowchart ou comparaison), cohérent avec l'illustration-titre, mêmes charte et style, libellés fournis explicitement en français, très peu de fioritures. - Libellés de schéma: des groupes nominaux courts et lisibles seuls (par exemple "Exemples de bonnes réponses"), pas des phrases verbales qui sonnent comme des ordres ("Montrer de bonnes réponses"). Quelques mots par libellé. Attention aux titres longs: Nano Banana rend fiablement les libellés courts mais peut faire une faute sur un titre long; la technique en deux temps limite le risque. Si le rendu d'un titre long reste incertain, indique à l'utilisateur qu'il peut demander à Gemini d'utiliser Nano Banana Pro, plus fiable pour le texte long. Gabarit, illustration-titre seule (exemple avec le titre 1): ``` Génère une image en respectant cette charte graphique: [bloc Charte graphique]. Contexte pour comprendre le sujet, à NE PAS afficher dans l'image: [ici le texte complet de la pastille]. Illustration-titre pour une pastille pédagogique sur les LLM. Composition épurée et moderne, focus graphique central iconique: un nuage de lettres et de mots qui se condense en une bulle de dialogue, évoquant un modèle de langage qui transforme du texte en réponse. Ce n'est pas un schéma de processus. Seul texte à afficher dans l'image: le titre, en en-tête, sans faute d'orthographe, en police sans serif corporate, sans sous-titre ni texte secondaire. Le titre exact: "TITRE EXACT". Tout le texte affiché est en français. Format 16:9. ``` Gabarit, illustration-titre plus schéma (deux images): ``` Prépares-toi à générer deux images séparées, dans deux fichiers distincts. Les deux partagent cette charte graphique: [bloc Charte graphique]. Tout texte affiché dans les images est en français. Contexte pour comprendre le sujet, à NE PAS afficher dans les images: [ici le texte complet de la pastille]. Image 1, illustration-titre: composition épurée et moderne, focus graphique central iconique qui illustre le sujet. Ce n'est pas un schéma de processus. Seul texte à afficher: le titre, en en-tête, sans faute d'orthographe, police sans serif corporate, sans sous-titre ni texte secondaire. Le titre exact: "TITRE EXACT". Format 16:9. Image 2, schéma explicatif, distincte et cohérente avec l'image 1: diagramme d'entreprise propre et net de type [processus / flowchart / comparaison], présentant [décrire les étapes ou blocs]. Libellés exacts à afficher en français: [liste des libellés]. Très peu de fioritures, focus sur la clarté, n'inclus pas le titre. Génère dans un premier temps seulement la première image (l'illustration-titre), et attends les instructions de l'utilisateur pour générer la 2e image. ``` ## Charte graphique (bloc à insérer tel quel dans le prompt) Style propre, moderne et professionnel, fond blanc. Palette: orange vif et bleu corporate foncé en dominantes, blanc et gris pour les respirations, accents multiculturels discrets (rouge, vert, jaune, bleu) reprenant subtilement un motif de couleurs de logo. Superpositions graphiques épurées: motifs géométriques abstraits, quartiers de cercle, lignes claires. Composition aérée, jamais surchargée. Typographie sans serif, propre et corporate. ## Format de sortie Le travail des sous-agents reste dans leur propre contexte, seuls leurs retours te reviennent. N'affiche donc que le livrable final (déjà corrigé par l'étape de revue), la conversation principale reste propre. Si l'utilisateur demande à voir l'atelier, restitue les cinq brouillons reçus et les constats de revue. Livrable final, dans cet ordre: - Quand une revue a eu lieu (systématiquement à la première génération), un court résumé "Ce que la revue a corrigé" (2 à 4 lignes) présentant les principaux ajustements, avant le reste. Lors d'ajustements ultérieurs sans revue, présente simplement la version ajustée sans ce résumé. - Le texte de la pastille (3 à 4 paragraphes), dans sa version corrigée. - Un bloc de code intitulé "Prompt images (à coller dans Gemini)", contenant le prompt unique prêt à copier. - Une courte section "Sources" listant 2 à 4 références principales issues de l'étape de recherche, de préférence officielles ou originales. Cette section sert à la vérification et n'a pas vocation à être publiée dans la pastille. Termine toujours par cette question exacte: "Comment trouvez-vous le texte, l'image titre et le diagramme (si généré) ? Si une partie vous semble trop complexe, ou si vous souhaitez affiner le focus d'un visuel pour qu'il soit encore plus épuré, n'hésitez pas à me le faire savoir, et je l'ajusterai." -> le personnage doit incarner le professeur un peu foufou qui allusine et obsédé par l'ai
015
Grok Image
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
507522
Nano Banana
Steps: 20, Sampler: DPM++ 2M Karras, CFG scale: 7, Seed: 413127992, Size: 1280x1600, Model hash: 7fa552ffb0, Model: nijidiffusedmix_v40, Denoising strength: 0.3, Clip skip: 2, ADetailer model: face_yolov8s.pt, ADetailer prompt: "dark blue eyes, freckles below eyes", ADetailer confidence: 0.3, ADetailer dilate erode: 4, ADetailer mask blur: 4, ADetailer denoising strength: 0.5, ADetailer inpaint only masked: True, ADetailer inpaint padding: 32, ADetailer version: 23.12.0, Version: v1.6.1
10493
Stable Diffusion
((abstract illusionism)), 8k full wide focus, (Masterpiece:1.2), ((surreal)),cinematic still, ((photorealistism)), (psychedelic imagery), trending on artstation, sharp focus, studio photo, intricate details, highly detailed, by greg rutkowski In the ethereal realm of the surreal dreamscape, envision a mesmerizing painting that beckons viewers into a kaleidoscope of fantastical wonders, a realm where imagination knows no bounds Negative prompt: nsfw, nude,blurry, low quality, oddly shaped, disfigured,unclear || undefined || mutations || deformities || off center || poor_composition Steps: 100, Sampler: DPM++ 3M SDE, CFG scale: 14.0, Seed: 33, Face restoration: CodeFormer, Size: 1280x832, Model hash: 1895a00bfc, Model: SSD-1B-A1111, Denoising strength: 0.31000000000000005, Mask blur: 4, Noise multiplier: 1.05, Version: v1.6.0-2-g4afaaf8a
0132
Stable Diffusion
Among the growing list of AI photo editing tools, Deepfake Maker takes a slightly different approach. While it’s best known for facial swaps and animated transformations, it also includes a surprisingly capable AI clothing removal feature. It’s not for everyone, but it’s an interesting choice for users looking for more dramatic or playful results. If you're after a quick way to generate bold clothing edits or outfit swaps, Deepfake Maker might surprise you. It’s not the most refined tool out there, but it’s definitely one of the most entertaining. 4. PxBee PxBee Source: PxBee Available On: Web Browser Price: Free *Annual subscription starts at $00.27/image and you can choose to download 50, 125, or 750 images monthly per year Pros Decent auto-detection for basic clothing removal Integrated with other tools Cons Not suitable for high-end projects Results take longer than expected to generate PxBee is better known for its stock photo collection and image generation features, but it also offers a basic AI clothing removal function built into its photo editing toolkit. While it's not as feature-rich or advanced as some dedicated AI editors, its simplicity and accessibility make it a decent option for casual users who want fast, no-frills results. PxBee may not be the most powerful AI clothing remover out there, but it's a solid pick for quick edits and simple concept visuals. Its intuitive UI and browser-based access make it appealing for users who want fast results without the learning curve. If you're after something basic, affordable, and easy to use, it's a tool worth exploring. 5. AI Face Swap AI Face Swap Source: AI Face Swap Available On: Web Browser Price: $1 per image Pros No manual masking needed Prompt-based customization Presets for inspiration Cons No fine-tuning or editing control With AI Face Swap, you can enter prompts like "red bikini", "formal suit", or even "nude", and the AI will generate a realistic version of your uploaded image that reflects your description. You can also choose from quick preset styles like bikini, romper, or skirt for faster results. This level of control makes the tool especially appealing to users looking for stylized edits, fantasy concepts, or mockup visuals. It also eliminates the need for manual masking or precise editing. 6. Bylo AI BYLO AI Source: Bylo.ai Available On: Web Browser Price: Free *Premium subscription available from $9.99 Pros Prompt-based clothing edits Free trial credits Part of a growing creative suite Cons No manual adjustment tools Limited free usage Relies heavily on prompts Bylo AI is a user-friendly AI platform that offers a range of creative features, including an AI Clothes Remover. The tool is built with simplicity in mind: just upload a photo, enter your prompt, and let the AI generate a clean, edited version. Bylo AI’s Clothes Remover is a great choice if you're looking for a lightweight, affordable tool that lets you experiment with clothing changes using just a text prompt. It won’t replace professional retouching tools, but for concept edits, social media content, or fun transformations, it delivers exactly what most casual users need. FAQs About AI Clothing Remover & Replacer 1. How do I remove clothes in photos? 2. What is the best online tool to remove clothes? 3. Are AI clothes removers safe?
31317
Hero
{ "prompt_type": "descriptive_replication", "reference_adherence": "STRICT_VISUAL_FIDELITY", "aspect_ratio": "3:4", "identity_lock": { "priority": "NONE", "instruction": "New female model. Completely different facial structure and overall appearance from any previous version. No resemblance to prior identity." }, "scene_constraints": { "people_count": "exactly one person in the frame", "single_subject_only": true, "no_other_people_visible": true, "no_extra_reflections": true, "no_background_figures": true, "full_body_visible": true, "head_to_toe_in_frame": true }, "subject": { "demographics": "young woman (age 20+), blonde", "presence": "only one visible person in the entire scene", "hair": { "color": "warm blonde (no dark roots)", "style": "side part, long straight hair", "texture": "smooth, natural fall, lightly voluminous" }, "face": { "structure": "distinct oval face shape with higher cheekbones and a slightly sharper jawline", "expression": "calm alluring expression with a subtle closed-lip smirk", "eyes": "almond-shaped eyes with a slightly lifted outer corner", "gaze": "looking at the phone screen through the mirror with composed confidence", "nose": "straight narrow nose with refined bridge", "lips": "naturally fuller lips with soft definition", "makeup": "minimal natural makeup, soft matte skin finish, muted rose lips, softly blended eyeshadow", "visibility": "partially covered by phone but clearly a different facial identity than any previous version" } }, "body": { "pose": "standing in front of mirror with one arm raised above head, legs naturally extended", "posture": "relaxed confident stance with slight hip shift", "anatomy": { "curves": "natural soft curves with realistic proportions", "chest": "bikini-supported, natural shape", "details": "visible collarbones, subtle shoulder definition, natural leg proportions" }, "feet": { "visibility": "fully visible at bottom of frame", "position": "flat on floor, natural stance" }, "skin_texture": "natural skin texture, no airbrushing" } }, "attire_and_accessories": { "clothing": { "top": "black triangle string bikini top with thin straps", "bottom_detail": "matching bikini bottom strings visible at hips" }, "footwear": "barefoot", "jewelry": { "necklaces": [], "wrists": [], "fingers": "no rings" }, "props": { "phone": "white iPhone (iPhone Pro style), held in right hand" } }, "environment": { "setting": "modern luxury bathroom interior, daytime", "foreground_element": { "mirror": "large rectangular mirror with GREY MARBLE FRAME", "material": "polished stone marble with subtle veining" }, "background": { "elements": [ "large marble wall tiles", "floating vanity", "minimalist sink and chrome fixtures", "folded white towels", "clean tiled floor fully visible" ], "style": "minimalist and clean", "clutter": "none" } }, "lighting_and_atmosphere": { "source": "natural daylight from camera-left", "quality": "soft directional light", "effects": [ "subtle highlights on shoulder and collarbone", "gentle shadow falloff along legs and feet" ], "contrast": "moderate, natural" }, "camera_and_technical": { "perspective": "mirror selfie", "camera_position": "eye level, slightly angled downward", "framing": "vertical 3:4, full head-to-toe body visible without cropping feet", "focus": "sharp on subject, background slightly softer", "visual_fidelity": "real smartphone camera look, mild natural noise" }, "realism_constraints": { "allowed": [ "natural skin texture", "minor asymmetry", "realistic lighting" ], "forbidden": [ "multiple people", "extra person", "background people", "beauty filters", "airbrushed skin", "anime style", "cartoon style", "cropped feet", "cut-off legs" ] } }
1356
Nano Banana
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
3246
Nano Banana
{ "template_name": "MARRLONN2027_V64_TERMINAL_CONDITIONAL_IMAGE3_REAL_HUMAN_MASTER", "template_type": "ABSOLUTE_STRICT_GOVERNANCE_MULTI_GENDER_DYNAMIC_IMAGE3_TEMPLATE", "version": "V64_TERMINAL_SEALED_DECISION_MATRIX_MICROSKIN_CANON_EOS_R50_KLING3_READY", "core_mission": "Create one final master still image with absolute IMAGE1 face identity, biological skin truth, age, beard, eyes and human presence; absolute IMAGE2 body anatomy, body scale, anatomical limits, physical structure and body-to-clothing logic; and intelligent conditional IMAGE3 transfer based only on the main visible pose-driving subject in IMAGE3. If IMAGE3 main subject is visually male for transfer purposes, clone pose, wardrobe, clothing, outfit, background, image style, lighting, accessories, worn objects, carried objects and visible scene objects when coherent, while preserving IMAGE1 identity and IMAGE2 body anatomy. If IMAGE3 main subject is visually female for transfer purposes, clone pose, pose logic, framing, composition, background, lighting, image style, camera language, scene mood and environment only, but never clone clothing, wardrobe, accessories, jewelry, makeup, hairstyle, shoes attached to the subject, or any object worn, attached to, held by, or carried by the female subject. If IMAGE3 has no human subject, multiple ambiguous subjects, or unclear transfer category, copy only scene, camera language, lighting, background, environment and non-conflicting objects while preserving IMAGE2 wardrobe and body logic. The final result must feel like a real imperfect human photographed with a Canon EOS R50, not AI, CGI, beauty retouch, plastic skin, waxy skin, luxury ad, actor replacement, model replacement, generic human replacement, gender-contaminated body transfer, or generated poster. The still image must be stable for Kling 3.0 video generation from 3 to 6 seconds.", "absolute_governance_principle": "IMAGE1_IDENTITY_SKIN_LOCK + IMAGE2_BODY_ANATOMY_LOCK + IMAGE3_MAIN_SUBJECT_CLASSIFICATION + IMAGE3_TRANSFER_DECISION_MATRIX + CONDITIONAL_WARDROBE_OBJECT_TRANSFER + REAL_REFERENCE_PRIORITY + MICROSKIN_REALISM + ZERO_BEAUTIFICATION + ZERO_GENERIC_REPLACEMENT + ZERO_PLASTIC_SKIN + ZERO_GENERATED_SCENE_FINISH + ANTI_POSE_DEFORMATION + STYLE_FIREWALL + CANON_EOS_R50_REAL_CAPTURE + KLING3_STABILITY", "primary_failure_to_prevent": "A recognizable subject that still feels AI-generated, too clean, too polished, too glossy, too actor-like, too model-like, too luxury-ad-like, gender-contaminated, anatomically forced, plastic, waxy, smooth, falsely hyperreal, or not biologically alive.", "success_definition": "The viewer must believe this is a real imperfect human physically present in a real environment, captured by a real Canon EOS R50 camera, with preserved IMAGE1 identity and skin truth, preserved IMAGE2 body reality, correct IMAGE3 transfer decision, visible pores, micro-wrinkles, natural asymmetry, believable hands, realistic clothing, imperfect scene texture, and Kling 3.0 stability.", "global_execution_hierarchy": [ "1_IMAGE1_FACE_IDENTITY_SKIN_AGE_BEARD_EYES", "2_IMAGE2_BODY_ANATOMY_SCALE_LIMITS", "3_IMAGE3_MAIN_SUBJECT_CLASSIFICATION_FOR_TRANSFER_ONLY", "4_IMAGE3_TRANSFER_DECISION_MATRIX", "5_CONDITIONAL_WARDROBE_ACCESSORY_OBJECT_TRANSFER", "6_REAL_REFERENCE_PIXEL_STRUCTURE_PRIORITY", "7_FINAL_REAL_HUMAN_PERCEPTION", "8_MICROSKIN_REALISM", "9_STYLE_FIREWALL_AND_ANTI_BEAUTY_TRANSFER", "10_ANTI_POSE_BODY_DEFORMATION", "11_ANTI_GENERATED_SCENE_FINISH", "12_CANON_EOS_R50_REAL_CAMERA_CAPTURE", "13_KLING3_VIDEO_STABILITY" ], "input_contract": { "IMAGE1": { "role": "ABSOLUTE_FACE_SKIN_IDENTITY_MASTER", "authority": "MAXIMUM_NON_NEGOTIABLE", "locked_domains": [ "facial_identity", "face_geometry", "eyes", "eyelids", "eyebrows", "nose", "mouth", "jaw", "cheeks", "chin", "forehead", "ears_when_visible", "neck_transition", "skin_tone", "individual_visible_pores", "irregular_pore_size", "skin_microtexture", "micro_wrinkles_around_eyes", "micro_wrinkles_around_mouth", "under_eye_texture", "subtle_redness", "small_blemishes", "minor_discoloration", "natural_asymmetry", "age_markers", "facial_fatigue", "beard_density", "beard_shape", "gray_beard_pattern", "mustache_texture", "random_beard_direction", "hairline", "real_biological_imperfection", "lived_in_human_presence" ], "mandatory_rules": [ "IMAGE1 is law, not inspiration", "transfer real skin truth from IMAGE1, never recreate a cleaner version", "preserve individual pores, irregular skin texture, micro-wrinkles, subtle redness and natural imperfections", "never improve attractiveness", "never normalize skin", "never clean the face", "never replace the subject with a generic realistic human", "never transfer identity, face geometry, age, skin style or beauty style from IMAGE3", "identity preservation overrides pose, wardrobe, style, scene, aesthetics and hyperrealism" ], "forbidden": [ "identity_drift", "generic_human_replacement", "celebrity_transfer", "actor_replacement", "model_replacement", "face_redesign", "beautification", "symmetry_correction", "youthification", "facial_slimming", "jaw_enhancement", "nose_refinement", "eye_enhancement", "lip_enhancement", "skin_smoothing", "skin_polishing", "skin_normalization", "makeup_effect", "cosmetic_cleanup", "plastic_skin", "waxy_skin", "porcelain_skin", "glass_skin", "AI_face", "perfect_idealized_face" ] }, "IMAGE2": { "role": "ABSOLUTE_BODY_ANATOMY_SCALE_MASTER", "authority": "MAXIMUM_NON_NEGOTIABLE", "locked_domains": [ "body_mass", "body_proportions", "shoulders", "torso", "waist", "arms", "forearms", "hands", "wrists", "legs", "stance", "scale", "natural_body_weight", "real_posture_limitations", "body_density", "real_anatomical_weight", "base_physical_structure", "body_to_clothing_fit_logic" ], "mandatory_rules": [ "IMAGE2 controls body anatomy, body mass, body scale and anatomical limits", "pose must adapt to IMAGE2 body", "wardrobe must fit IMAGE2 body", "body must never adapt to an idealized pose, fashion silhouette or actor posture", "hands must remain anatomically believable and stable for Kling 3.0", "clothing cannot create a different body", "if IMAGE3 wardrobe conflicts with IMAGE2 anatomy, adapt wardrobe to IMAGE2 instead of changing body" ], "forbidden": [ "slimming", "muscle_invention", "hero_body_transformation", "fashion_body_stylization", "incorrect_anatomy", "long_arms", "short_arms", "deformed_hands", "wrong_body_scale", "catalog_body_correction", "fake_torso_taper", "mannequin_posture", "body_redesign_to_fit_IMAGE3" ] }, "IMAGE3": { "role": "CONDITIONAL_DYNAMIC_REFERENCE", "authority": "CONDITIONAL_BY_MAIN_VISIBLE_POSE_DRIVING_SUBJECT", "universal_allowed": [ "pose", "pose_logic", "camera_angle", "framing", "composition", "background", "environment", "scene_mood", "lighting_direction", "image_style", "apparent_camera_language", "vehicles", "architecture", "people_in_background", "birds", "non_conflicting_visible_scene_objects" ], "universal_forbidden": [ "face_contamination", "skin_contamination", "identity_override", "facial_geometry_override", "beauty_style_transfer", "cinematic_glamour_applied_to_face", "makeup_transfer", "skin_tone_replacement", "facial_expression_beautification", "celebrity_face_borrowing", "actor_charisma_transfer", "luxury_catalog_facial_treatment", "body_redesign_to_match_IMAGE3" ] } }, "image3_subject_classification_gate": { "enabled": true, "mandatory": true, "purpose": "Classify IMAGE3 only for transfer logic, never as personal identity judgment.", "primary_subject_detection": "Detect the main visible pose-driving subject in IMAGE3: the person most central to pose, framing, scale and composition.", "classification_basis": "Use visible presentation, wardrobe context, pose-driving role and scene function only for transfer logic; never infer personal identity, protected identity or real-life gender identity.", "classification_options": [ "visual_male_transfer_subject", "visual_female_transfer_subject", "multiple_subjects", "no_human_subject", "uncertain_or_ambiguous" ], "if_multiple_subjects": { "rule": "Select only the main pose-driving subject. Secondary people remain background or environment context only.", "do_not_copy": [ "wardrobe_from_secondary_people", "accessories_from_secondary_people", "objects_carried_by_secondary_people", "identity_or_skin_traits_from_any_IMAGE3_person" ] }, "if_no_human_subject": { "copy_from_IMAGE3": [ "background", "environment", "scene_mood", "lighting_direction", "camera_angle", "composition", "image_style", "vehicles", "architecture", "non_conflicting_objects" ], "keep_from_IMAGE2": [ "body_anatomy", "pose_if_no_user_pose_instruction_exists", "wardrobe_unless_explicitly_changed_by_user" ] }, "if_uncertain_or_ambiguous": { "rule": "Protect IMAGE1 and IMAGE2 first. Copy only pose, scene, lighting, background, camera language and non-gendered environment.", "do_not_guess_gendered_wardrobe": true, "do_not_copy_gendered_accessories": true, "do_not_copy_worn_attached_or_carried_objects": true } }, "image3_transfer_decision_matrix": { "enabled": true, "mandatory": true, "visual_male_transfer_subject": { "copy": [ "pose", "pose_logic", "body_orientation_translated_to_IMAGE2", "wardrobe", "clothing", "outfit", "fabric_style", "shoes_when_visible", "accessories", "neutral_or_masculine_jewelry", "watch", "glasses", "hat", "objects_held_by_subject", "objects_worn_by_subject", "objects_attached_to_subject", "background", "lighting", "image_style", "camera_language", "scene_mood", "environment", "vehicles", "architecture", "non_conflicting_scene_objects" ], "filters": [ "must_fit_IMAGE2_body", "must_not_change_IMAGE1_identity", "must_not_transfer_IMAGE3_skin_or_face", "must_not_deform_hands", "must_not_create_catalog_body", "must_not_create_CGI_fashion_render", "must_remain_Kling_stable" ] }, "visual_female_transfer_subject": { "copy": [ "pose", "pose_logic", "body_orientation_translated_to_IMAGE2", "camera_angle", "framing", "composition", "background", "lighting", "image_style", "camera_language", "scene_mood", "environment", "vehicles", "architecture", "people_in_background", "birds", "non_attached_scene_props", "non_conflicting_environmental_objects" ], "block": [ "female_clothing", "female_wardrobe", "female_outfit", "female_vestments", "female_shoes_attached_to_subject", "female_accessories", "female_jewelry", "earrings", "female_bracelets", "female_necklaces", "female_rings", "female_watch_if_gender_coded", "female_makeup", "female_hairstyle", "female_body_traits", "female_glamour_cues", "handbag", "purse", "neckline", "cleavage_styling", "female_fabric_drape", "female_body_fitting_silhouette", "female_jewelry_dependent_hand_pose", "any_wearable_item", "any_carried_item", "any_handheld_item", "any_attached_item", "any_body_dependent_item", "any_gender_coded_item", "any_fashion_dependent_item_from_female_subject" ], "keep_from_IMAGE2": [ "final_clothing", "final_wardrobe", "body_anatomy", "body_proportions", "body_scale", "base_body_to_clothing_logic" ] }, "multiple_subjects": { "rule": "Use only main pose-driving subject for transfer. Secondary people stay as background context.", "block_secondary_subject_transfer": true }, "no_human_subject": { "copy": [ "scene", "camera_language", "lighting", "background", "environment", "objects", "vehicles", "architecture", "image_style" ], "keep_from_IMAGE2": [ "body", "wardrobe", "pose_unless_user_explicitly_requests_pose_change" ] }, "uncertain_or_ambiguous": { "copy": [ "pose_if_clear", "scene", "lighting", "background", "camera_language", "environment" ], "keep_from_IMAGE2": [ "body", "wardrobe", "presentation" ], "block": [ "gendered_clothing", "gendered_accessories", "worn_objects", "attached_objects", "carried_objects" ] } }, "image3_gender_conditional_transfer_system": { "enabled": true, "mandatory": true, "if_IMAGE3_subject_is_male": { "copy_from_IMAGE3": [ "pose", "pose_logic", "body_orientation_translated_intelligently", "camera_angle", "framing", "composition", "background", "environment", "scene_mood", "lighting_direction", "image_style", "apparent_camera_language", "wardrobe", "clothing", "outfit", "fabric_style", "shoes_when_visible", "accessories", "jewelry_if_masculine_or_neutral", "watch", "glasses", "hat", "objects_held_by_the_male_subject", "objects_worn_by_the_male_subject", "objects_attached_to_the_male_subject", "vehicles", "architecture", "people_in_background", "birds", "non_conflicting_visible_scene_objects" ], "adaptation_rules": [ "clone male IMAGE3 wardrobe and accessories when visually coherent", "fit IMAGE3 wardrobe to IMAGE2 body anatomy and proportions", "translate IMAGE3 pose through IMAGE2 body limits", "preserve IMAGE2 body mass and scale even when copying IMAGE3 outfit", "do not reshape IMAGE2 body to force IMAGE3 wardrobe", "do not alter IMAGE1 face, skin, beard, age, eyes, or identity", "objects and accessories must remain physically plausible and naturally integrated", "if male IMAGE3 wardrobe conflicts with IMAGE2 anatomy, adapt wardrobe instead of changing body" ], "object_accessory_risk_filter": { "block_if": [ "covers_or_deforms_face", "breaks_hands", "forces_extra_fingers", "creates_glossy_AI_look", "creates_unstable_reflections", "creates_identity_drift", "creates_Kling_instability", "creates_scene_overload", "looks_like_CGI_prop" ] } }, "if_IMAGE3_subject_is_female": { "copy_from_IMAGE3": [ "pose", "pose_logic", "body_orientation_translated_intelligently", "camera_angle", "framing", "composition", "background", "environment", "scene_mood", "lighting_direction", "image_style", "apparent_camera_language", "vehicles", "architecture", "people_in_background", "birds", "non_attached_scene_props", "non_conflicting_environmental_objects" ], "keep_from_IMAGE2": [ "final_clothing", "final_wardrobe", "body_anatomy", "body_proportions", "body_scale", "presentation_consistency", "base_body_to_clothing_logic" ], "strict_rule": "If IMAGE3 subject is visually female for transfer purposes, copy pose, pose logic, scene, background, lighting, image style, camera style, environment and non-attached scene elements only. Do not copy any wearable, carried, handheld, attached, body-dependent, gender-coded, or fashion-dependent item from the female subject." }, "universal_rules": [ "IMAGE1 always controls face, identity, skin, beard, age and human truth", "IMAGE2 always controls body anatomy, body scale and anatomical limits", "IMAGE3 never controls identity or skin", "pose must be anatomically translated through IMAGE2", "final result must remain real human, imperfect, non-plastic, non-beautified and Canon EOS R50 believable" ] }, "domain_separation_system": { "identity_domain": { "source": "IMAGE1", "influence": 1.0, "mode": "ABSOLUTE_NON_NEGOTIABLE" }, "skin_domain": { "source": "IMAGE1", "influence": 1.0, "mode": "TRANSFER_NOT_RECREATE", "forbid_cleaner_version": true, "forbid_fake_procedural_texture": true, "forbid_skin_reinterpretation": true }, "body_domain": { "source": "IMAGE2", "influence": 1.0, "mode": "ABSOLUTE_NON_NEGOTIABLE" }, "wardrobe_domain": { "source": "CONDITIONAL", "influence": "CONTROLLED", "mode": "IMAGE3_IF_MALE__IMAGE2_IF_IMAGE3_FEMALE_NO_HUMAN_UNCERTAIN_OR_INCOMPATIBLE", "rules": [ "if IMAGE3 main subject is male, copy wardrobe, clothing, outfit, accessories and worn/carried objects when coherent", "if IMAGE3 main subject is female, do not copy wardrobe, clothing, accessories, jewelry, makeup, hairstyle, or worn/carried/attached objects", "if IMAGE3 has no human subject, preserve IMAGE2 wardrobe unless user explicitly requests a change", "if transfer category is uncertain, preserve IMAGE2 wardrobe and copy only pose, scene, lighting and camera language", "wardrobe must always fit IMAGE2 body anatomy", "wardrobe must never alter IMAGE1 identity or IMAGE2 body structure" ] }, "pose_domain": { "source": "IMAGE3", "influence": "CONTROLLED", "mode": "TRANSLATE_THROUGH_IMAGE2_BODY_LIMITS" }, "object_domain": { "source": "CONDITIONAL", "influence": "FILTERED", "mode": "MALE_HELD_WORN_ATTACHED_OBJECTS_ALLOWED_IF_COHERENT__FEMALE_ONLY_ENVIRONMENT_OBJECTS", "rules": [ "if IMAGE3 subject is male, copy held, worn, attached and environmental objects when coherent and stable", "if IMAGE3 subject is female, copy only environmental objects not worn, attached to, or carried by the female subject", "objects must not change identity, body, hands or human realism", "objects must remain stable for Kling 3.0" ] }, "style_domain": { "source": "IMAGE3", "influence": "CONTROLLED", "allowed": [ "color_mood", "camera_angle", "framing", "composition", "contrast_level", "lighting_direction", "texture_of_capture", "photo_or_video_language" ], "forbidden": [ "beauty_skin_style", "glamour_facial_lighting", "plastic_highlights", "fake_HDR", "poster_render_finish", "CGI_material_style", "skin_smoothing_style", "luxury_ad_polish_on_face" ] }, "lighting_domain": { "source": "IMAGE3", "influence": "DIRECTION_AND_ENVIRONMENT_ONLY", "mode": "FACE_SKIN_RESPONSE_FROM_IMAGE1", "forbid_beauty_lighting_on_face": true, "forbid_skin_washout": true, "forbid_glamour_fill": true, "forbid_fake_cinematic_skin_softening": true } }, "real_reference_pixel_priority": { "enabled": true, "mandatory": true, "priority": "DOMINANT", "purpose": "Reduce generated-looking output by preserving real visual structure from IMAGE1 and IMAGE2 whenever possible.", "rules": [ "prioritize real facial structure from IMAGE1 instead of rebuilding a new face", "prioritize real skin texture logic from IMAGE1 instead of inventing new skin", "prioritize real body geometry from IMAGE2 instead of reconstructing a synthetic body", "preserve real asymmetry, age markers, facial fatigue and imperfect human cues", "do not over-generate areas that already work in IMAGE1 or IMAGE2", "perform controlled adaptation, not total reinterpretation", "never generate a new subject when the real reference structure can be preserved" ], "forbidden": [ "full_synthetic_facial_reconstruction", "new_skin_generated_over_IMAGE1_truth", "generic_realistic_face_replacement", "body_redesign_from_scratch", "AI_cleaned_version_of_subject", "reference_ignored_in_favor_of_prompt_aesthetics" ] }, "micro_skin_realism_system": { "enabled": true, "mandatory": true, "source": "IMAGE1", "requirements": [ "individual_visible_pores", "irregular_pore_size", "skin_microtexture", "micro_wrinkles_around_eyes", "micro_wrinkles_around_mouth", "forehead_texture", "under_eye_texture", "subtle_redness", "minor_blemishes", "tiny_discolorations", "natural_asymmetry", "real_beard_root_texture", "non_uniform_cheek_texture", "non_uniform_forehead_texture", "real_face_to_neck_transition", "small_real_imperfections_inherited_from_IMAGE1" ], "strict_forbidden": [ "airbrushed_skin", "beauty_skin", "cosmetic_cleanup", "pore_removal", "fake_pore_overlay", "procedural_pore_stamping", "over_sharpened_pore_texture", "plastic_highlights", "waxy_subsurface", "porcelain_effect", "smooth_cheek_gradient", "AI_skin_texture", "skin_made_cleaner_than_IMAGE1", "foundation_look" ], "specular_policy": { "allowed": [ "minimal_natural_human_skin_reflectance", "subtle_non_cosmetic_skin_response_only_when_physically_plausible" ], "forbidden": [ "invented_sweat_drops", "visible_oil_effect_unless_explicitly_requested", "cosmetic_shine", "beauty_gloss", "wet_render_skin", "uniform_shiny_face", "makeup_glow", "plastic_shine" ] } }, "humanity_over_hyperrealism_rule": { "enabled": true, "mandatory": true, "priority": "RAW_HUMANITY_OVER_DECORATIVE_HYPERREALISM", "rule": "Human realism is more important than hyperreal detail. If hyperrealism creates plastic skin, synthetic lighting, luxury render finish, CGI texture, poster aesthetics, over-detail, false sharpness, or unnatural perfection, reduce hyperrealism and preserve raw photographic humanity.", "reject_if": [ "hyperreal_detail_looks_synthetic", "skin_detail_looks_procedural", "lighting_looks_like_render", "scene_looks_like_luxury_ad", "human_subject_looks_too_perfect", "camera_look_feels_fake" ] }, "anti_pose_body_deformation_system": { "enabled": true, "mandatory": true, "rules": [ "clone IMAGE3 pose but translate it through IMAGE2 body anatomy", "do not break shoulders", "do not elongate arms", "do not shrink torso", "do not slim waist", "do not force feminine posture when IMAGE3 subject is female", "do not force hands into object-dependent female accessory poses", "reduce pose accuracy if anatomy, identity, hands or Kling stability are at risk" ], "reject_if": [ "pose_changes_body_mass", "pose_deforms_hands", "pose_breaks_shoulders", "pose_feminizes_subject_when_not_desired", "pose_creates_mannequin_posture", "pose_creates_unstable_Kling_geometry" ] }, "anti_generated_scene_finish": { "enabled": true, "mandatory": true, "forbidden": [ "perfect_mansion_render", "overly_clean_background", "repetitive_stone_texture", "over_glossy_car_paint", "catalog_car_reflections", "impossible_chrome_reflections", "luxury_ad_finish", "polished_showroom_look", "fake_cinematic_poster_lighting", "CGI_environment", "AI_background_blend", "overprocessed_clarity", "perfect_global_sharpness", "plastic_material_response", "scene_cleaner_than_reality" ], "requirements": [ "scene_must_feel_physically_real", "background_must_contain_natural_imperfections", "materials_must_have_plausible_real_world_response", "contact_shadows_must_be_believable", "subject_must_integrate_naturally_into_scene", "environment_realism_must_support_human_subject_not_overpower_it" ] }, "real_human_perception_system": { "enabled": true, "mandatory": true, "requirements": [ "biologically_alive_presence", "real_facial_gravity", "organic_facial_tension", "calm_natural_expression", "subtle_facial_asymmetry", "real_skin_imperfection", "ordinary_human_texture", "realistic_face_neck_transition", "real_body_weight", "real_clothing_tension", "non_mannequin_posture", "real_interaction_with_objects" ], "reject_if": [ "subject_feels_synthetic", "subject_feels_generated", "subject_feels_too_clean", "skin_looks_perfect", "eyes_look_synthetic", "expression_feels_frozen", "face_looks_corrected", "humanity_is_achieved_by_changing_identity" ] }, "body_hands_wardrobe_system": { "enabled": true, "body_source": "IMAGE2", "wardrobe_source": "CONDITIONAL_IMAGE3_MALE_OR_IMAGE2_FEMALE_NO_HUMAN_UNCERTAIN_CONFLICT", "pose_source": "IMAGE3_CONTROLLED", "hands_rules": [ "five_fingers", "natural_hand_proportion", "real_knuckles", "real_fingernail_structure", "subtle_finger_wrinkles", "natural_finger_curvature", "correct_wrist_connection", "real_contact_pressure", "hands_must_not_look_flat", "hands_must_not_look_waxy", "hands_must_not_align_unnaturally", "no_floating_hands", "no_fused_fingers", "no_extra_fingers", "no_mannequin_hands" ], "wardrobe_rules": [ "if_IMAGE3_subject_is_male_wardrobe_clothing_outfit_and_coherent_accessories_may_be_copied_from_IMAGE3", "if_IMAGE3_subject_is_female_wardrobe_must_follow_IMAGE2_and_female_clothing_accessories_from_IMAGE3_are_blocked", "if_IMAGE3_has_no_human_subject_or_is_ambiguous_preserve_IMAGE2_wardrobe_unless_user_explicitly_requests_otherwise", "wardrobe_must_always_fit_IMAGE2_body_anatomy", "clothing_must_not_reshape_shoulders", "clothing_must_not_slim_torso", "fabric_tension_must_follow_real_body_mass", "fabric_must_show_real_texture_natural_folds_and_non_perfect_wear" ] }, "camera_and_rendering_system": { "camera_model": "Canon EOS R50", "capture_intent": "real unretouched camera photograph", "look": "natural Canon EOS R50 real-world capture", "rules": [ "natural_color", "balanced_exposure", "no_blown_highlights", "plausible_optics", "realistic_depth_of_field", "natural_shadow_falloff", "real_microcontrast", "no_over_sharpening", "no_over_grading", "no_fake_HDR", "no_beauty_grade", "real_material_response", "no_CGI_render_finish", "no_computational_skin_smoothing" ], "lens_policy": { "portrait_closeup": "85mm portrait feeling only when composition is close-up and appropriate", "full_body_or_vehicle_scene": "use natural Canon EOS R50 perspective appropriate for the scene; do not force 85mm if it distorts composition" } }, "kling3_preparation": { "ready_for_video": true, "duration": "3_to_6_seconds", "must_survive": [ "identity_continuity", "skin_continuity", "eye_stability", "beard_edge_stability", "hand_integrity", "clothing_stability", "geometry_stability", "subject_background_integration", "zero_shimmer", "zero_warp", "stable_object_contact", "stable_face_outline", "stable_neck_and_collar", "stable_material_reflections" ], "forbidden": [ "latent_AI_skin", "synthetic_eyes", "overpolished_face", "ambiguous_facial_texture", "unstable_hands", "unclear_edges", "flicker_prone_glossy_face", "AI_edge_halos", "clothing_seams_likely_to_crawl", "glossy_zones_likely_to_flicker" ] }, "validation_gate": { "must_pass": [ "same_person_as_IMAGE1", "same_skin_truth_as_IMAGE1", "same_body_reality_as_IMAGE2", "correct_IMAGE3_main_subject_classification_for_transfer_only", "correct_conditional_IMAGE3_transfer", "if_IMAGE3_is_male_coherent_wardrobe_clothing_accessories_and_objects_are_copied_when_physically_plausible", "if_IMAGE3_is_female_clothing_wardrobe_accessories_jewelry_makeup_hairstyle_and_worn_carried_attached_objects_are_blocked", "if_IMAGE3_has_no_human_or_is_uncertain_IMAGE2_wardrobe_body_are_preserved_while_scene_camera_lights_are_copied", "no_AI_skin", "no_plastic_skin", "no_waxy_skin", "no_makeup_invention", "no_generic_handsome_face", "no_beauty_retouch", "credible_real_human_perception", "Canon_EOS_R50_real_camera_feeling", "usable_as_master_frame_for_Kling3" ], "reject_if": [ "looks_beautified", "looks_cleaner_than_IMAGE1", "skin_tone_more_uniform_than_IMAGE1", "beard_looks_cleaner_than_IMAGE1", "looks_like_different_person", "feels_synthetic", "style_from_IMAGE3_contaminates_skin", "male_IMAGE3_clothing_not_copied_when_coherent", "female_IMAGE3_wardrobe_or_accessories_copied", "female_IMAGE3_carried_worn_handheld_attached_or_body_dependent_objects_copied", "accessory_breaks_hands_or_identity", "object_unstable_for_Kling", "scene_looks_CGI", "lighting_looks_poster_render", "pose_deforms_body", "wardrobe_changes_body_mass", "unstable_for_video" ] }, "render_priority": [ "FACE_IDENTITY_FROM_IMAGE1", "BODY_ANATOMY_FROM_IMAGE2", "IMAGE3_MAIN_SUBJECT_CLASSIFICATION_FOR_TRANSFER_ONLY", "IMAGE3_TRANSFER_DECISION_MATRIX", "CONDITIONAL_IMAGE3_TRANSFER_BY_SUBJECT_TYPE", "REAL_REFERENCE_STRUCTURE_PRIORITY", "SKIN_TRUTH_MICROTEXTURE_FROM_IMAGE1", "FINAL_HUMAN_PERCEPTION", "MALE_IMAGE3_WARDROBE_ACCESSORIES_OBJECTS_WHEN_COHERENT", "FEMALE_IMAGE3_SCENE_POSE_LIGHT_STYLE_ONLY_WITH_WARDROBE_ACCESSORIES_OBJECTS_BLOCKED", "NO_HUMAN_OR_UNCERTAIN_IMAGE3_SCENE_CAMERA_LIGHT_ONLY", "ANTI_POSE_BODY_DEFORMATION", "STYLE_FIREWALL", "ANTI_GENERATED_SCENE_FINISH", "CANON_EOS_R50_REALISM", "KLING3_STABILITY" ], "final_rule": "Reject immediately if the result looks beautified, polished, synthetic, glamorous, too clean, too smooth, actor-like, model-like, luxury-ad-like, mannequin-like, CGI-like, plastic, waxy, oily, makeup-covered, showroom-rendered, poster-like, gender-contaminated, anatomically forced, falsely hyperreal, or AI-generated. Accept only if IMAGE1 identity and micro-skin truth remain intact, IMAGE2 body anatomy and scale remain intact, IMAGE3 transfer follows main-subject classification and decision-matrix logic, male IMAGE3 wardrobe/accessories/objects transfer when coherent, female IMAGE3 wardrobe/accessories/worn-carried-attached objects are fully blocked, real reference structure is preserved, scene avoids generated perfection, and final output feels like a real imperfect human captured by a Canon EOS R50 and stable enough for Kling 3.0." }
1139
Nano Banana
Ultra-realistic 8K scene of the person in the photo, stepping out of and breaking the panel border of a giant open manga book lying on the floor of an aesthetic bedroom. The comic pages depict a surreal 4-panel black-and-white manga with screentones, bold ink lines, and authentic Japanese layout, featuring the manga version of the same person trapped inside. The larger, real-life version emerges dynamically into the comic’s surface, vibrant, colorful, and photorealistic, in sharp contrast to the monochrome manga world. Both versions interact directly, creating a meta, humorous encounter. Highly detailed, crisp manga line art seamlessly integrated with vivid, realistic
24964
Nano Banana
gett nude Learn to create better images with AI! Check out our crash course in prompt engineering & AI art generation! AM1970 posted about 2 hours ago 8 views 0 comments Behold this realistic masterpiece: a slim, elegant woman with fiery red hair tied in a long ponytail, defined eyebrows framing her striking blue eyes, and pale skin dotted with charming freckles across her face and chest. Captured from a side angle, she radiates happiness with a closed-mouth smile, her upturned nose adding to her allure. Dressed in a vibrant, colorful fabric robe pieced together with zips, buttons, seams, and intricate stitching—fully clothed yet teasing partial nudity with subtle cleavage and teardrop-shaped, saggy breasts for a natural, confident vibe. Illuminated by dramatic rim lighting that highlights every textured, detailed inch of her perfect skin, this high-quality image is a celebration of beauty, realism, and artistry! Read more... Comments (0) Add a comment Stunning handsome readhead woman by AM1970 • Follow 0 realistic, masterpiece, high quality BREAK 1woman, 28 y. o. slim fit woman, red hair, long pony tail, defined eyebrows, (fully clothed) , pale skin, (realistic skin:1.3) | (textured skin:1.3) | (detailed skin:1.3) | (perfect skin:1.3) , (upturned nose) , freckles on face and chest, colorful fabric combining to form robe zips, buttons, seams, stitching), partial nudity, teardrop boobs, blue eyes, (cleavage:1.2) , saggy breasts. side angle. happy| closed mouth. ((rim lighting) ) Negative prompt: unrealistic, saturated, high contrast, big nose, painting, drawing, sketch, cartoon, anime, manga, render, CG, 3d, watermark, signature, label, nsfw, nudity, fetish, latex, pvc, suggestive pose, weapon, combat, blood, war scene, aggressive expression, harsh shadows, heavy makeup, color tint, saturation, chromatic aberration, lens flare, low resolution, blur, noise, artifacts, over-retouching, plastic skin, oversharpening, watermark, text, logo Generation parameters 832x1216 Model used CyberRealistic XL v7.0 [Download AI model] More by AM1970 Twitter Instagram Say hi 👋 PromptHero is built by ElevenYellow . Navigation Create AI images Academy - Learn AI PromptHero Community Affiliate program Openjourney AI Challenges Verified profiles Plans Advertise FAQ 🟢 Status Learn AI Content Creation with AI for Brands and Products Midjourney for graphic design & art professionals Crash course in generative AI & prompt engineering for images AI influencers and consistent characters Create custom AI models and LoRas by fine-tuning Stable Diffusion Master your composition: advanced AI image generation with ControlNet Guides Best AI Upscalers 2025 Best AI Video Editors 2025 Best AI 3D Models 2025 Best AI Image Editors 2025 Best AI Notetakers 2025 The Best AI Prompts Best AI Prompts Best Veo Prompts Best Hailuo Prompts Best Wanx Prompts Best Hunyuan Prompts Best Seedance Prompts Best Pixverse Prompts Best Luma Prompts Best Kling Prompts Best Runway Prompts Best Nano Banana Prompts Best SeedEdit Prompts Best Qwen Prompts Best FLUX Prompts Best Recraft Prompts Best Ideogram Prompts Best Seedream 4 Prompts Best Google Imagen Prompts Best ChatGPT Image Prompts Best Stable Diffusion Prompts Best DreamShaper Prompts Best Realistic Vision Prompts Best epiCRealism Prompts Best majicMIX realistic Prompts Best Midjourney Prompts Best Openjourney Prompts Best ChilloutMix Prompts Best Deliberate Prompts Best ReV Animated Prompts Best Perfect World Prompts Best Dreamlike Photoreal Prompts Best Dreamlike Diffusion Prompts Best Anything Prompts Best Protogen Prompts Best AbyssOrangeMix Prompts Best Grapefruit Prompts Best Kenshi Prompts Best Analog Diffusion Prompts Best SynthwavePunk Prompts Best Illuminati Diffusion Prompts Best Arcane Diffusion Prompts Best Mo-Di Diffusion Prompts Best Redshift Diffusion Prompts Best Waifu Diffusion Prompts Best TrinArt Prompts Best Poolsuite Diffusion Prompts Best Funko Diffusion Prompts Best Text to Pokémon Prompts Best DALL-E Prompts Best Fashion Diffusion Prompts AI Models All AI models Image generation AI Models Video generation AI Models Large Language Models LoRA Textual Inversion ControlNet Hypernetwork Aesthetic Gradient LyCORIS VAE Poses Wildcards Workflows Other Other sites Dreamshot Jobician 2025 © PromptHero @prompthero
1362
FLUX
Among the growing list of AI photo editing tools, Deepfake Maker takes a slightly different approach. While it’s best known for facial swaps and animated transformations, it also includes a surprisingly capable AI clothing removal feature. It’s not for everyone, but it’s an interesting choice for users looking for more dramatic or playful results. If you're after a quick way to generate bold clothing edits or outfit swaps, Deepfake Maker might surprise you. It’s not the most refined tool out there, but it’s definitely one of the most entertaining. 4. PxBee PxBee Source: PxBee Available On: Web Browser Price: Free *Annual subscription starts at $00.27/image and you can choose to download 50, 125, or 750 images monthly per year Pros Decent auto-detection for basic clothing removal Integrated with other tools Cons Not suitable for high-end projects Results take longer than expected to generate PxBee is better known for its stock photo collection and image generation features, but it also offers a basic AI clothing removal function built into its photo editing toolkit. While it's not as feature-rich or advanced as some dedicated AI editors, its simplicity and accessibility make it a decent option for casual users who want fast, no-frills results. PxBee may not be the most powerful AI clothing remover out there, but it's a solid pick for quick edits and simple concept visuals. Its intuitive UI and browser-based access make it appealing for users who want fast results without the learning curve. If you're after something basic, affordable, and easy to use, it's a tool worth exploring. 5. AI Face Swap AI Face Swap Source: AI Face Swap Available On: Web Browser Price: $1 per image Pros No manual masking needed Prompt-based customization Presets for inspiration Cons No fine-tuning or editing control With AI Face Swap, you can enter prompts like "red bikini", "formal suit", or even "nude", and the AI will generate a realistic version of your uploaded image that reflects your description. You can also choose from quick preset styles like bikini, romper, or skirt for faster results. This level of control makes the tool especially appealing to users looking for stylized edits, fantasy concepts, or mockup visuals. It also eliminates the need for manual masking or precise editing. 6. Bylo AI BYLO AI Source: Bylo.ai Available On: Web Browser Price: Free *Premium subscription available from $9.99 Pros Prompt-based clothing edits Free trial credits Part of a growing creative suite Cons No manual adjustment tools Limited free usage Relies heavily on prompts Bylo AI is a user-friendly AI platform that offers a range of creative features, including an AI Clothes Remover. The tool is built with simplicity in mind: just upload a photo, enter your prompt, and let the AI generate a clean, edited version. Bylo AI’s Clothes Remover is a great choice if you're looking for a lightweight, affordable tool that lets you experiment with clothing changes using just a text prompt. It won’t replace professional retouching tools, but for concept edits, social media content, or fun transformations, it delivers exactly what most casual users need. FAQs About AI Clothing Remover & Replacer 1. How do I remove clothes in photos? 2. What is the best online tool to remove clothes? 3. Are AI clothes removers safe?
42502
Hero
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
5393
Nano Banana
High-definition quality, photo version. Shot with a Canon EOS 5D Mark IV and an EF 100mm f/2.8L Macro IS USM lens, f/4.5, 1/200s, ISO 400. Best quality, ultra HD, 8K, award-winning photography. --ar 3:4 High-definition, realistic skin textures, extremely detailed facial textures, cinematic lighting, and a natural atmosphere, no cartoon style.
10353
Sora
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
727622
Z Image
((abstract illusionism)), 8k full wide focus, (Masterpiece:1.2), ((surreal)),cinematic still, ((photorealistism)), (psychedelic imagery), trending on artstation, sharp focus, studio photo, intricate details, highly detailed, by greg rutkowski In the ethereal realm of the surreal dreamscape, envision a mesmerizing painting that beckons viewers into a kaleidoscope of fantastical wonders, a realm where imagination knows no bounds Negative prompt: nsfw, nude,blurry, low quality, oddly shaped, disfigured,unclear || undefined || mutations || deformities || off center || poor_composition Steps: 100, Sampler: DPM++ 3M SDE, CFG scale: 14.0, Seed: 33, Face restoration: CodeFormer, Size: 1280x832, Model hash: 1895a00bfc, Model: SSD-1B-A1111, Denoising strength: 0.31000000000000005, Mask blur: 4, Noise multiplier: 1.05, Version: v1.6.0-2-g4afaaf8a
0132
Stable Diffusion
Among the growing list of AI photo editing tools, Deepfake Maker takes a slightly different approach. While it’s best known for facial swaps and animated transformations, it also includes a surprisingly capable AI clothing removal feature. It’s not for everyone, but it’s an interesting choice for users looking for more dramatic or playful results. If you're after a quick way to generate bold clothing edits or outfit swaps, Deepfake Maker might surprise you. It’s not the most refined tool out there, but it’s definitely one of the most entertaining. 4. PxBee PxBee Source: PxBee Available On: Web Browser Price: Free *Annual subscription starts at $00.27/image and you can choose to download 50, 125, or 750 images monthly per year Pros Decent auto-detection for basic clothing removal Integrated with other tools Cons Not suitable for high-end projects Results take longer than expected to generate PxBee is better known for its stock photo collection and image generation features, but it also offers a basic AI clothing removal function built into its photo editing toolkit. While it's not as feature-rich or advanced as some dedicated AI editors, its simplicity and accessibility make it a decent option for casual users who want fast, no-frills results. PxBee may not be the most powerful AI clothing remover out there, but it's a solid pick for quick edits and simple concept visuals. Its intuitive UI and browser-based access make it appealing for users who want fast results without the learning curve. If you're after something basic, affordable, and easy to use, it's a tool worth exploring. 5. AI Face Swap AI Face Swap Source: AI Face Swap Available On: Web Browser Price: $1 per image Pros No manual masking needed Prompt-based customization Presets for inspiration Cons No fine-tuning or editing control With AI Face Swap, you can enter prompts like "red bikini", "formal suit", or even "nude", and the AI will generate a realistic version of your uploaded image that reflects your description. You can also choose from quick preset styles like bikini, romper, or skirt for faster results. This level of control makes the tool especially appealing to users looking for stylized edits, fantasy concepts, or mockup visuals. It also eliminates the need for manual masking or precise editing. 6. Bylo AI BYLO AI Source: Bylo.ai Available On: Web Browser Price: Free *Premium subscription available from $9.99 Pros Prompt-based clothing edits Free trial credits Part of a growing creative suite Cons No manual adjustment tools Limited free usage Relies heavily on prompts Bylo AI is a user-friendly AI platform that offers a range of creative features, including an AI Clothes Remover. The tool is built with simplicity in mind: just upload a photo, enter your prompt, and let the AI generate a clean, edited version. Bylo AI’s Clothes Remover is a great choice if you're looking for a lightweight, affordable tool that lets you experiment with clothing changes using just a text prompt. It won’t replace professional retouching tools, but for concept edits, social media content, or fun transformations, it delivers exactly what most casual users need. FAQs About AI Clothing Remover & Replacer 1. How do I remove clothes in photos? 2. What is the best online tool to remove clothes? 3. Are AI clothes removers safe?
31317
Hero
{ "generation_request": { "meta_data": { "tool": "NanoBanana Pro", "task_type": "photobooth_contact_sheet_fashion_grid", "version": "v1.1_REFERENCE_IDENTITY_GRID_EN", "priority": "highest", "language": "en" }, "input": { "mode": "image_to_image", "reference_image_usage": "required", "notes": "Use the uploaded reference image to preserve the same person’s face, body features, hair, makeup and clothing. Maintain identity consistency in all frames." }, "output_settings": { "aspect_ratio": "4:5", "orientation": "portrait", "resolution_target": "ultra_high_res", "num_images": 1, "layout": { "type": "grid", "rows": 4, "cols": 4, "gutter": "thin_black_borders", "outer_border": "black_frame" }, "sharpness": "slightly_soft_flash", "film_grain": "medium_analog", "color_grade": "warm_2000s_flash_tone", "dynamic_range": "natural_not_hdr" }, "creative_prompt": { "scene_summary": "Early 2000s disposable camera photobooth contact sheet featuring the same person from the reference image in all 16 frames. Playful, candid, flash-lit aesthetic.", "environment": { "background": "plain white wall indoors", "lighting": "direct on-camera flash, slight hard shadow behind subject, subtle hotspot on skin, authentic point-and-shoot look" }, "subject": { "identity_consistency": "exact same person as reference image in all frames", "hair": "match reference image exactly", "makeup": "match reference image exactly", "wardrobe": "preserve the exact clothing from the reference image with accurate fabric texture and fit", "accessories": "keep accessories consistent with reference image" }, "poses_and_expressions": [ "hands on hips confident pose", "soft pout toward camera", "gentle smile", "hand on chin thoughtful look", "playful wink", "hands in hair", "slight shoulder turn", "looking off to the side", "serious model face", "soft laugh", "blowing a kiss", "tilted head cute expression", "crossed arms relaxed pose", "side glance over shoulder", "chin resting on hand", "natural relaxed expression" ], "style": { "aesthetic": "2000s photobooth disposable camera", "texture": "visible film grain, dust particles, tiny scratches, mild vignetting", "color": "slightly warm faded tones, subtle flash flattening", "lens": "35mm point-and-shoot softness" } }, "negative_prompt": [ "text", "logos", "watermark", "different outfit between frames", "different hairstyle", "different person", "face distortion", "identity drift", "overly polished editorial look", "HDR" ] } }
391
Ultra-realistic 8K scene of the person in the photo, stepping out of and breaking the panel border of a giant open manga book lying on the floor of an aesthetic bedroom. The comic pages depict a surreal 4-panel black-and-white manga with screentones, bold ink lines, and authentic Japanese layout, featuring the manga version of the same person trapped inside. The larger, real-life version emerges dynamically into the comic’s surface, vibrant, colorful, and photorealistic, in sharp contrast to the monochrome manga world. Both versions interact directly, creating a meta, humorous encounter. Highly detailed, crisp manga line art seamlessly integrated with vivid, realistic
432772
Nano Banana
Among the growing list of AI photo editing tools, Deepfake Maker takes a slightly different approach. While it’s best known for facial swaps and animated transformations, it also includes a surprisingly capable AI clothing removal feature. It’s not for everyone, but it’s an interesting choice for users looking for more dramatic or playful results. If you're after a quick way to generate bold clothing edits or outfit swaps, Deepfake Maker might surprise you. It’s not the most refined tool out there, but it’s definitely one of the most entertaining. 4. PxBee PxBee Source: PxBee Available On: Web Browser Price: Free *Annual subscription starts at $00.27/image and you can choose to download 50, 125, or 750 images monthly per year Pros Decent auto-detection for basic clothing removal Integrated with other tools Cons Not suitable for high-end projects Results take longer than expected to generate PxBee is better known for its stock photo collection and image generation features, but it also offers a basic AI clothing removal function built into its photo editing toolkit. While it's not as feature-rich or advanced as some dedicated AI editors, its simplicity and accessibility make it a decent option for casual users who want fast, no-frills results. PxBee may not be the most powerful AI clothing remover out there, but it's a solid pick for quick edits and simple concept visuals. Its intuitive UI and browser-based access make it appealing for users who want fast results without the learning curve. If you're after something basic, affordable, and easy to use, it's a tool worth exploring. 5. AI Face Swap AI Face Swap Source: AI Face Swap Available On: Web Browser Price: $1 per image Pros No manual masking needed Prompt-based customization Presets for inspiration Cons No fine-tuning or editing control With AI Face Swap, you can enter prompts like "red bikini", "formal suit", or even "nude", and the AI will generate a realistic version of your uploaded image that reflects your description. You can also choose from quick preset styles like bikini, romper, or skirt for faster results. This level of control makes the tool especially appealing to users looking for stylized edits, fantasy concepts, or mockup visuals. It also eliminates the need for manual masking or precise editing. 6. Bylo AI BYLO AI Source: Bylo.ai Available On: Web Browser Price: Free *Premium subscription available from $9.99 Pros Prompt-based clothing edits Free trial credits Part of a growing creative suite Cons No manual adjustment tools Limited free usage Relies heavily on prompts Bylo AI is a user-friendly AI platform that offers a range of creative features, including an AI Clothes Remover. The tool is built with simplicity in mind: just upload a photo, enter your prompt, and let the AI generate a clean, edited version. Bylo AI’s Clothes Remover is a great choice if you're looking for a lightweight, affordable tool that lets you experiment with clothing changes using just a text prompt. It won’t replace professional retouching tools, but for concept edits, social media content, or fun transformations, it delivers exactly what most casual users need. FAQs About AI Clothing Remover & Replacer 1. How do I remove clothes in photos? 2. What is the best online tool to remove clothes? 3. Are AI clothes removers safe?
42502
Hero
Among the growing list of AI photo editing tools, Deepfake Maker takes a slightly different approach. While it’s best known for facial swaps and animated transformations, it also includes a surprisingly capable AI clothing removal feature. It’s not for everyone, but it’s an interesting choice for users looking for more dramatic or playful results. If you're after a quick way to generate bold clothing edits or outfit swaps, Deepfake Maker might surprise you. It’s not the most refined tool out there, but it’s definitely one of the most entertaining. 4. PxBee PxBee Source: PxBee Available On: Web Browser Price: Free *Annual subscription starts at $00.27/image and you can choose to download 50, 125, or 750 images monthly per year Pros Decent auto-detection for basic clothing removal Integrated with other tools Cons Not suitable for high-end projects Results take longer than expected to generate PxBee is better known for its stock photo collection and image generation features, but it also offers a basic AI clothing removal function built into its photo editing toolkit. While it's not as feature-rich or advanced as some dedicated AI editors, its simplicity and accessibility make it a decent option for casual users who want fast, no-frills results. PxBee may not be the most powerful AI clothing remover out there, but it's a solid pick for quick edits and simple concept visuals. Its intuitive UI and browser-based access make it appealing for users who want fast results without the learning curve. If you're after something basic, affordable, and easy to use, it's a tool worth exploring. 5. AI Face Swap AI Face Swap Source: AI Face Swap Available On: Web Browser Price: $1 per image Pros No manual masking needed Prompt-based customization Presets for inspiration Cons No fine-tuning or editing control With AI Face Swap, you can enter prompts like "red bikini", "formal suit", or even "nude", and the AI will generate a realistic version of your uploaded image that reflects your description. You can also choose from quick preset styles like bikini, romper, or skirt for faster results. This level of control makes the tool especially appealing to users looking for stylized edits, fantasy concepts, or mockup visuals. It also eliminates the need for manual masking or precise editing. 6. Bylo AI BYLO AI Source: Bylo.ai Available On: Web Browser Price: Free *Premium subscription available from $9.99 Pros Prompt-based clothing edits Free trial credits Part of a growing creative suite Cons No manual adjustment tools Limited free usage Relies heavily on prompts Bylo AI is a user-friendly AI platform that offers a range of creative features, including an AI Clothes Remover. The tool is built with simplicity in mind: just upload a photo, enter your prompt, and let the AI generate a clean, edited version. Bylo AI’s Clothes Remover is a great choice if you're looking for a lightweight, affordable tool that lets you experiment with clothing changes using just a text prompt. It won’t replace professional retouching tools, but for concept edits, social media content, or fun transformations, it delivers exactly what most casual users need. FAQs About AI Clothing Remover & Replacer 1. How do I remove clothes in photos? 2. What is the best online tool to remove clothes? 3. Are AI clothes removers safe?
2677
Hero
{ "template_name": "MARRLONN2027_V64_TERMINAL_CONDITIONAL_IMAGE3_REAL_HUMAN_MASTER", "template_type": "ABSOLUTE_STRICT_GOVERNANCE_MULTI_GENDER_DYNAMIC_IMAGE3_TEMPLATE", "version": "V64_TERMINAL_SEALED_DECISION_MATRIX_MICROSKIN_CANON_EOS_R50_KLING3_READY", "core_mission": "Create one final master still image with absolute IMAGE1 face identity, biological skin truth, age, beard, eyes and human presence; absolute IMAGE2 body anatomy, body scale, anatomical limits, physical structure and body-to-clothing logic; and intelligent conditional IMAGE3 transfer based only on the main visible pose-driving subject in IMAGE3. If IMAGE3 main subject is visually male for transfer purposes, clone pose, wardrobe, clothing, outfit, background, image style, lighting, accessories, worn objects, carried objects and visible scene objects when coherent, while preserving IMAGE1 identity and IMAGE2 body anatomy. If IMAGE3 main subject is visually female for transfer purposes, clone pose, pose logic, framing, composition, background, lighting, image style, camera language, scene mood and environment only, but never clone clothing, wardrobe, accessories, jewelry, makeup, hairstyle, shoes attached to the subject, or any object worn, attached to, held by, or carried by the female subject. If IMAGE3 has no human subject, multiple ambiguous subjects, or unclear transfer category, copy only scene, camera language, lighting, background, environment and non-conflicting objects while preserving IMAGE2 wardrobe and body logic. The final result must feel like a real imperfect human photographed with a Canon EOS R50, not AI, CGI, beauty retouch, plastic skin, waxy skin, luxury ad, actor replacement, model replacement, generic human replacement, gender-contaminated body transfer, or generated poster. The still image must be stable for Kling 3.0 video generation from 3 to 6 seconds.", "absolute_governance_principle": "IMAGE1_IDENTITY_SKIN_LOCK + IMAGE2_BODY_ANATOMY_LOCK + IMAGE3_MAIN_SUBJECT_CLASSIFICATION + IMAGE3_TRANSFER_DECISION_MATRIX + CONDITIONAL_WARDROBE_OBJECT_TRANSFER + REAL_REFERENCE_PRIORITY + MICROSKIN_REALISM + ZERO_BEAUTIFICATION + ZERO_GENERIC_REPLACEMENT + ZERO_PLASTIC_SKIN + ZERO_GENERATED_SCENE_FINISH + ANTI_POSE_DEFORMATION + STYLE_FIREWALL + CANON_EOS_R50_REAL_CAPTURE + KLING3_STABILITY", "primary_failure_to_prevent": "A recognizable subject that still feels AI-generated, too clean, too polished, too glossy, too actor-like, too model-like, too luxury-ad-like, gender-contaminated, anatomically forced, plastic, waxy, smooth, falsely hyperreal, or not biologically alive.", "success_definition": "The viewer must believe this is a real imperfect human physically present in a real environment, captured by a real Canon EOS R50 camera, with preserved IMAGE1 identity and skin truth, preserved IMAGE2 body reality, correct IMAGE3 transfer decision, visible pores, micro-wrinkles, natural asymmetry, believable hands, realistic clothing, imperfect scene texture, and Kling 3.0 stability.", "global_execution_hierarchy": [ "1_IMAGE1_FACE_IDENTITY_SKIN_AGE_BEARD_EYES", "2_IMAGE2_BODY_ANATOMY_SCALE_LIMITS", "3_IMAGE3_MAIN_SUBJECT_CLASSIFICATION_FOR_TRANSFER_ONLY", "4_IMAGE3_TRANSFER_DECISION_MATRIX", "5_CONDITIONAL_WARDROBE_ACCESSORY_OBJECT_TRANSFER", "6_REAL_REFERENCE_PIXEL_STRUCTURE_PRIORITY", "7_FINAL_REAL_HUMAN_PERCEPTION", "8_MICROSKIN_REALISM", "9_STYLE_FIREWALL_AND_ANTI_BEAUTY_TRANSFER", "10_ANTI_POSE_BODY_DEFORMATION", "11_ANTI_GENERATED_SCENE_FINISH", "12_CANON_EOS_R50_REAL_CAMERA_CAPTURE", "13_KLING3_VIDEO_STABILITY" ], "input_contract": { "IMAGE1": { "role": "ABSOLUTE_FACE_SKIN_IDENTITY_MASTER", "authority": "MAXIMUM_NON_NEGOTIABLE", "locked_domains": [ "facial_identity", "face_geometry", "eyes", "eyelids", "eyebrows", "nose", "mouth", "jaw", "cheeks", "chin", "forehead", "ears_when_visible", "neck_transition", "skin_tone", "individual_visible_pores", "irregular_pore_size", "skin_microtexture", "micro_wrinkles_around_eyes", "micro_wrinkles_around_mouth", "under_eye_texture", "subtle_redness", "small_blemishes", "minor_discoloration", "natural_asymmetry", "age_markers", "facial_fatigue", "beard_density", "beard_shape", "gray_beard_pattern", "mustache_texture", "random_beard_direction", "hairline", "real_biological_imperfection", "lived_in_human_presence" ], "mandatory_rules": [ "IMAGE1 is law, not inspiration", "transfer real skin truth from IMAGE1, never recreate a cleaner version", "preserve individual pores, irregular skin texture, micro-wrinkles, subtle redness and natural imperfections", "never improve attractiveness", "never normalize skin", "never clean the face", "never replace the subject with a generic realistic human", "never transfer identity, face geometry, age, skin style or beauty style from IMAGE3", "identity preservation overrides pose, wardrobe, style, scene, aesthetics and hyperrealism" ], "forbidden": [ "identity_drift", "generic_human_replacement", "celebrity_transfer", "actor_replacement", "model_replacement", "face_redesign", "beautification", "symmetry_correction", "youthification", "facial_slimming", "jaw_enhancement", "nose_refinement", "eye_enhancement", "lip_enhancement", "skin_smoothing", "skin_polishing", "skin_normalization", "makeup_effect", "cosmetic_cleanup", "plastic_skin", "waxy_skin", "porcelain_skin", "glass_skin", "AI_face", "perfect_idealized_face" ] }, "IMAGE2": { "role": "ABSOLUTE_BODY_ANATOMY_SCALE_MASTER", "authority": "MAXIMUM_NON_NEGOTIABLE", "locked_domains": [ "body_mass", "body_proportions", "shoulders", "torso", "waist", "arms", "forearms", "hands", "wrists", "legs", "stance", "scale", "natural_body_weight", "real_posture_limitations", "body_density", "real_anatomical_weight", "base_physical_structure", "body_to_clothing_fit_logic" ], "mandatory_rules": [ "IMAGE2 controls body anatomy, body mass, body scale and anatomical limits", "pose must adapt to IMAGE2 body", "wardrobe must fit IMAGE2 body", "body must never adapt to an idealized pose, fashion silhouette or actor posture", "hands must remain anatomically believable and stable for Kling 3.0", "clothing cannot create a different body", "if IMAGE3 wardrobe conflicts with IMAGE2 anatomy, adapt wardrobe to IMAGE2 instead of changing body" ], "forbidden": [ "slimming", "muscle_invention", "hero_body_transformation", "fashion_body_stylization", "incorrect_anatomy", "long_arms", "short_arms", "deformed_hands", "wrong_body_scale", "catalog_body_correction", "fake_torso_taper", "mannequin_posture", "body_redesign_to_fit_IMAGE3" ] }, "IMAGE3": { "role": "CONDITIONAL_DYNAMIC_REFERENCE", "authority": "CONDITIONAL_BY_MAIN_VISIBLE_POSE_DRIVING_SUBJECT", "universal_allowed": [ "pose", "pose_logic", "camera_angle", "framing", "composition", "background", "environment", "scene_mood", "lighting_direction", "image_style", "apparent_camera_language", "vehicles", "architecture", "people_in_background", "birds", "non_conflicting_visible_scene_objects" ], "universal_forbidden": [ "face_contamination", "skin_contamination", "identity_override", "facial_geometry_override", "beauty_style_transfer", "cinematic_glamour_applied_to_face", "makeup_transfer", "skin_tone_replacement", "facial_expression_beautification", "celebrity_face_borrowing", "actor_charisma_transfer", "luxury_catalog_facial_treatment", "body_redesign_to_match_IMAGE3" ] } }, "image3_subject_classification_gate": { "enabled": true, "mandatory": true, "purpose": "Classify IMAGE3 only for transfer logic, never as personal identity judgment.", "primary_subject_detection": "Detect the main visible pose-driving subject in IMAGE3: the person most central to pose, framing, scale and composition.", "classification_basis": "Use visible presentation, wardrobe context, pose-driving role and scene function only for transfer logic; never infer personal identity, protected identity or real-life gender identity.", "classification_options": [ "visual_male_transfer_subject", "visual_female_transfer_subject", "multiple_subjects", "no_human_subject", "uncertain_or_ambiguous" ], "if_multiple_subjects": { "rule": "Select only the main pose-driving subject. Secondary people remain background or environment context only.", "do_not_copy": [ "wardrobe_from_secondary_people", "accessories_from_secondary_people", "objects_carried_by_secondary_people", "identity_or_skin_traits_from_any_IMAGE3_person" ] }, "if_no_human_subject": { "copy_from_IMAGE3": [ "background", "environment", "scene_mood", "lighting_direction", "camera_angle", "composition", "image_style", "vehicles", "architecture", "non_conflicting_objects" ], "keep_from_IMAGE2": [ "body_anatomy", "pose_if_no_user_pose_instruction_exists", "wardrobe_unless_explicitly_changed_by_user" ] }, "if_uncertain_or_ambiguous": { "rule": "Protect IMAGE1 and IMAGE2 first. Copy only pose, scene, lighting, background, camera language and non-gendered environment.", "do_not_guess_gendered_wardrobe": true, "do_not_copy_gendered_accessories": true, "do_not_copy_worn_attached_or_carried_objects": true } }, "image3_transfer_decision_matrix": { "enabled": true, "mandatory": true, "visual_male_transfer_subject": { "copy": [ "pose", "pose_logic", "body_orientation_translated_to_IMAGE2", "wardrobe", "clothing", "outfit", "fabric_style", "shoes_when_visible", "accessories", "neutral_or_masculine_jewelry", "watch", "glasses", "hat", "objects_held_by_subject", "objects_worn_by_subject", "objects_attached_to_subject", "background", "lighting", "image_style", "camera_language", "scene_mood", "environment", "vehicles", "architecture", "non_conflicting_scene_objects" ], "filters": [ "must_fit_IMAGE2_body", "must_not_change_IMAGE1_identity", "must_not_transfer_IMAGE3_skin_or_face", "must_not_deform_hands", "must_not_create_catalog_body", "must_not_create_CGI_fashion_render", "must_remain_Kling_stable" ] }, "visual_female_transfer_subject": { "copy": [ "pose", "pose_logic", "body_orientation_translated_to_IMAGE2", "camera_angle", "framing", "composition", "background", "lighting", "image_style", "camera_language", "scene_mood", "environment", "vehicles", "architecture", "people_in_background", "birds", "non_attached_scene_props", "non_conflicting_environmental_objects" ], "block": [ "female_clothing", "female_wardrobe", "female_outfit", "female_vestments", "female_shoes_attached_to_subject", "female_accessories", "female_jewelry", "earrings", "female_bracelets", "female_necklaces", "female_rings", "female_watch_if_gender_coded", "female_makeup", "female_hairstyle", "female_body_traits", "female_glamour_cues", "handbag", "purse", "neckline", "cleavage_styling", "female_fabric_drape", "female_body_fitting_silhouette", "female_jewelry_dependent_hand_pose", "any_wearable_item", "any_carried_item", "any_handheld_item", "any_attached_item", "any_body_dependent_item", "any_gender_coded_item", "any_fashion_dependent_item_from_female_subject" ], "keep_from_IMAGE2": [ "final_clothing", "final_wardrobe", "body_anatomy", "body_proportions", "body_scale", "base_body_to_clothing_logic" ] }, "multiple_subjects": { "rule": "Use only main pose-driving subject for transfer. Secondary people stay as background context.", "block_secondary_subject_transfer": true }, "no_human_subject": { "copy": [ "scene", "camera_language", "lighting", "background", "environment", "objects", "vehicles", "architecture", "image_style" ], "keep_from_IMAGE2": [ "body", "wardrobe", "pose_unless_user_explicitly_requests_pose_change" ] }, "uncertain_or_ambiguous": { "copy": [ "pose_if_clear", "scene", "lighting", "background", "camera_language", "environment" ], "keep_from_IMAGE2": [ "body", "wardrobe", "presentation" ], "block": [ "gendered_clothing", "gendered_accessories", "worn_objects", "attached_objects", "carried_objects" ] } }, "image3_gender_conditional_transfer_system": { "enabled": true, "mandatory": true, "if_IMAGE3_subject_is_male": { "copy_from_IMAGE3": [ "pose", "pose_logic", "body_orientation_translated_intelligently", "camera_angle", "framing", "composition", "background", "environment", "scene_mood", "lighting_direction", "image_style", "apparent_camera_language", "wardrobe", "clothing", "outfit", "fabric_style", "shoes_when_visible", "accessories", "jewelry_if_masculine_or_neutral", "watch", "glasses", "hat", "objects_held_by_the_male_subject", "objects_worn_by_the_male_subject", "objects_attached_to_the_male_subject", "vehicles", "architecture", "people_in_background", "birds", "non_conflicting_visible_scene_objects" ], "adaptation_rules": [ "clone male IMAGE3 wardrobe and accessories when visually coherent", "fit IMAGE3 wardrobe to IMAGE2 body anatomy and proportions", "translate IMAGE3 pose through IMAGE2 body limits", "preserve IMAGE2 body mass and scale even when copying IMAGE3 outfit", "do not reshape IMAGE2 body to force IMAGE3 wardrobe", "do not alter IMAGE1 face, skin, beard, age, eyes, or identity", "objects and accessories must remain physically plausible and naturally integrated", "if male IMAGE3 wardrobe conflicts with IMAGE2 anatomy, adapt wardrobe instead of changing body" ], "object_accessory_risk_filter": { "block_if": [ "covers_or_deforms_face", "breaks_hands", "forces_extra_fingers", "creates_glossy_AI_look", "creates_unstable_reflections", "creates_identity_drift", "creates_Kling_instability", "creates_scene_overload", "looks_like_CGI_prop" ] } }, "if_IMAGE3_subject_is_female": { "copy_from_IMAGE3": [ "pose", "pose_logic", "body_orientation_translated_intelligently", "camera_angle", "framing", "composition", "background", "environment", "scene_mood", "lighting_direction", "image_style", "apparent_camera_language", "vehicles", "architecture", "people_in_background", "birds", "non_attached_scene_props", "non_conflicting_environmental_objects" ], "keep_from_IMAGE2": [ "final_clothing", "final_wardrobe", "body_anatomy", "body_proportions", "body_scale", "presentation_consistency", "base_body_to_clothing_logic" ], "strict_rule": "If IMAGE3 subject is visually female for transfer purposes, copy pose, pose logic, scene, background, lighting, image style, camera style, environment and non-attached scene elements only. Do not copy any wearable, carried, handheld, attached, body-dependent, gender-coded, or fashion-dependent item from the female subject." }, "universal_rules": [ "IMAGE1 always controls face, identity, skin, beard, age and human truth", "IMAGE2 always controls body anatomy, body scale and anatomical limits", "IMAGE3 never controls identity or skin", "pose must be anatomically translated through IMAGE2", "final result must remain real human, imperfect, non-plastic, non-beautified and Canon EOS R50 believable" ] }, "domain_separation_system": { "identity_domain": { "source": "IMAGE1", "influence": 1.0, "mode": "ABSOLUTE_NON_NEGOTIABLE" }, "skin_domain": { "source": "IMAGE1", "influence": 1.0, "mode": "TRANSFER_NOT_RECREATE", "forbid_cleaner_version": true, "forbid_fake_procedural_texture": true, "forbid_skin_reinterpretation": true }, "body_domain": { "source": "IMAGE2", "influence": 1.0, "mode": "ABSOLUTE_NON_NEGOTIABLE" }, "wardrobe_domain": { "source": "CONDITIONAL", "influence": "CONTROLLED", "mode": "IMAGE3_IF_MALE__IMAGE2_IF_IMAGE3_FEMALE_NO_HUMAN_UNCERTAIN_OR_INCOMPATIBLE", "rules": [ "if IMAGE3 main subject is male, copy wardrobe, clothing, outfit, accessories and worn/carried objects when coherent", "if IMAGE3 main subject is female, do not copy wardrobe, clothing, accessories, jewelry, makeup, hairstyle, or worn/carried/attached objects", "if IMAGE3 has no human subject, preserve IMAGE2 wardrobe unless user explicitly requests a change", "if transfer category is uncertain, preserve IMAGE2 wardrobe and copy only pose, scene, lighting and camera language", "wardrobe must always fit IMAGE2 body anatomy", "wardrobe must never alter IMAGE1 identity or IMAGE2 body structure" ] }, "pose_domain": { "source": "IMAGE3", "influence": "CONTROLLED", "mode": "TRANSLATE_THROUGH_IMAGE2_BODY_LIMITS" }, "object_domain": { "source": "CONDITIONAL", "influence": "FILTERED", "mode": "MALE_HELD_WORN_ATTACHED_OBJECTS_ALLOWED_IF_COHERENT__FEMALE_ONLY_ENVIRONMENT_OBJECTS", "rules": [ "if IMAGE3 subject is male, copy held, worn, attached and environmental objects when coherent and stable", "if IMAGE3 subject is female, copy only environmental objects not worn, attached to, or carried by the female subject", "objects must not change identity, body, hands or human realism", "objects must remain stable for Kling 3.0" ] }, "style_domain": { "source": "IMAGE3", "influence": "CONTROLLED", "allowed": [ "color_mood", "camera_angle", "framing", "composition", "contrast_level", "lighting_direction", "texture_of_capture", "photo_or_video_language" ], "forbidden": [ "beauty_skin_style", "glamour_facial_lighting", "plastic_highlights", "fake_HDR", "poster_render_finish", "CGI_material_style", "skin_smoothing_style", "luxury_ad_polish_on_face" ] }, "lighting_domain": { "source": "IMAGE3", "influence": "DIRECTION_AND_ENVIRONMENT_ONLY", "mode": "FACE_SKIN_RESPONSE_FROM_IMAGE1", "forbid_beauty_lighting_on_face": true, "forbid_skin_washout": true, "forbid_glamour_fill": true, "forbid_fake_cinematic_skin_softening": true } }, "real_reference_pixel_priority": { "enabled": true, "mandatory": true, "priority": "DOMINANT", "purpose": "Reduce generated-looking output by preserving real visual structure from IMAGE1 and IMAGE2 whenever possible.", "rules": [ "prioritize real facial structure from IMAGE1 instead of rebuilding a new face", "prioritize real skin texture logic from IMAGE1 instead of inventing new skin", "prioritize real body geometry from IMAGE2 instead of reconstructing a synthetic body", "preserve real asymmetry, age markers, facial fatigue and imperfect human cues", "do not over-generate areas that already work in IMAGE1 or IMAGE2", "perform controlled adaptation, not total reinterpretation", "never generate a new subject when the real reference structure can be preserved" ], "forbidden": [ "full_synthetic_facial_reconstruction", "new_skin_generated_over_IMAGE1_truth", "generic_realistic_face_replacement", "body_redesign_from_scratch", "AI_cleaned_version_of_subject", "reference_ignored_in_favor_of_prompt_aesthetics" ] }, "micro_skin_realism_system": { "enabled": true, "mandatory": true, "source": "IMAGE1", "requirements": [ "individual_visible_pores", "irregular_pore_size", "skin_microtexture", "micro_wrinkles_around_eyes", "micro_wrinkles_around_mouth", "forehead_texture", "under_eye_texture", "subtle_redness", "minor_blemishes", "tiny_discolorations", "natural_asymmetry", "real_beard_root_texture", "non_uniform_cheek_texture", "non_uniform_forehead_texture", "real_face_to_neck_transition", "small_real_imperfections_inherited_from_IMAGE1" ], "strict_forbidden": [ "airbrushed_skin", "beauty_skin", "cosmetic_cleanup", "pore_removal", "fake_pore_overlay", "procedural_pore_stamping", "over_sharpened_pore_texture", "plastic_highlights", "waxy_subsurface", "porcelain_effect", "smooth_cheek_gradient", "AI_skin_texture", "skin_made_cleaner_than_IMAGE1", "foundation_look" ], "specular_policy": { "allowed": [ "minimal_natural_human_skin_reflectance", "subtle_non_cosmetic_skin_response_only_when_physically_plausible" ], "forbidden": [ "invented_sweat_drops", "visible_oil_effect_unless_explicitly_requested", "cosmetic_shine", "beauty_gloss", "wet_render_skin", "uniform_shiny_face", "makeup_glow", "plastic_shine" ] } }, "humanity_over_hyperrealism_rule": { "enabled": true, "mandatory": true, "priority": "RAW_HUMANITY_OVER_DECORATIVE_HYPERREALISM", "rule": "Human realism is more important than hyperreal detail. If hyperrealism creates plastic skin, synthetic lighting, luxury render finish, CGI texture, poster aesthetics, over-detail, false sharpness, or unnatural perfection, reduce hyperrealism and preserve raw photographic humanity.", "reject_if": [ "hyperreal_detail_looks_synthetic", "skin_detail_looks_procedural", "lighting_looks_like_render", "scene_looks_like_luxury_ad", "human_subject_looks_too_perfect", "camera_look_feels_fake" ] }, "anti_pose_body_deformation_system": { "enabled": true, "mandatory": true, "rules": [ "clone IMAGE3 pose but translate it through IMAGE2 body anatomy", "do not break shoulders", "do not elongate arms", "do not shrink torso", "do not slim waist", "do not force feminine posture when IMAGE3 subject is female", "do not force hands into object-dependent female accessory poses", "reduce pose accuracy if anatomy, identity, hands or Kling stability are at risk" ], "reject_if": [ "pose_changes_body_mass", "pose_deforms_hands", "pose_breaks_shoulders", "pose_feminizes_subject_when_not_desired", "pose_creates_mannequin_posture", "pose_creates_unstable_Kling_geometry" ] }, "anti_generated_scene_finish": { "enabled": true, "mandatory": true, "forbidden": [ "perfect_mansion_render", "overly_clean_background", "repetitive_stone_texture", "over_glossy_car_paint", "catalog_car_reflections", "impossible_chrome_reflections", "luxury_ad_finish", "polished_showroom_look", "fake_cinematic_poster_lighting", "CGI_environment", "AI_background_blend", "overprocessed_clarity", "perfect_global_sharpness", "plastic_material_response", "scene_cleaner_than_reality" ], "requirements": [ "scene_must_feel_physically_real", "background_must_contain_natural_imperfections", "materials_must_have_plausible_real_world_response", "contact_shadows_must_be_believable", "subject_must_integrate_naturally_into_scene", "environment_realism_must_support_human_subject_not_overpower_it" ] }, "real_human_perception_system": { "enabled": true, "mandatory": true, "requirements": [ "biologically_alive_presence", "real_facial_gravity", "organic_facial_tension", "calm_natural_expression", "subtle_facial_asymmetry", "real_skin_imperfection", "ordinary_human_texture", "realistic_face_neck_transition", "real_body_weight", "real_clothing_tension", "non_mannequin_posture", "real_interaction_with_objects" ], "reject_if": [ "subject_feels_synthetic", "subject_feels_generated", "subject_feels_too_clean", "skin_looks_perfect", "eyes_look_synthetic", "expression_feels_frozen", "face_looks_corrected", "humanity_is_achieved_by_changing_identity" ] }, "body_hands_wardrobe_system": { "enabled": true, "body_source": "IMAGE2", "wardrobe_source": "CONDITIONAL_IMAGE3_MALE_OR_IMAGE2_FEMALE_NO_HUMAN_UNCERTAIN_CONFLICT", "pose_source": "IMAGE3_CONTROLLED", "hands_rules": [ "five_fingers", "natural_hand_proportion", "real_knuckles", "real_fingernail_structure", "subtle_finger_wrinkles", "natural_finger_curvature", "correct_wrist_connection", "real_contact_pressure", "hands_must_not_look_flat", "hands_must_not_look_waxy", "hands_must_not_align_unnaturally", "no_floating_hands", "no_fused_fingers", "no_extra_fingers", "no_mannequin_hands" ], "wardrobe_rules": [ "if_IMAGE3_subject_is_male_wardrobe_clothing_outfit_and_coherent_accessories_may_be_copied_from_IMAGE3", "if_IMAGE3_subject_is_female_wardrobe_must_follow_IMAGE2_and_female_clothing_accessories_from_IMAGE3_are_blocked", "if_IMAGE3_has_no_human_subject_or_is_ambiguous_preserve_IMAGE2_wardrobe_unless_user_explicitly_requests_otherwise", "wardrobe_must_always_fit_IMAGE2_body_anatomy", "clothing_must_not_reshape_shoulders", "clothing_must_not_slim_torso", "fabric_tension_must_follow_real_body_mass", "fabric_must_show_real_texture_natural_folds_and_non_perfect_wear" ] }, "camera_and_rendering_system": { "camera_model": "Canon EOS R50", "capture_intent": "real unretouched camera photograph", "look": "natural Canon EOS R50 real-world capture", "rules": [ "natural_color", "balanced_exposure", "no_blown_highlights", "plausible_optics", "realistic_depth_of_field", "natural_shadow_falloff", "real_microcontrast", "no_over_sharpening", "no_over_grading", "no_fake_HDR", "no_beauty_grade", "real_material_response", "no_CGI_render_finish", "no_computational_skin_smoothing" ], "lens_policy": { "portrait_closeup": "85mm portrait feeling only when composition is close-up and appropriate", "full_body_or_vehicle_scene": "use natural Canon EOS R50 perspective appropriate for the scene; do not force 85mm if it distorts composition" } }, "kling3_preparation": { "ready_for_video": true, "duration": "3_to_6_seconds", "must_survive": [ "identity_continuity", "skin_continuity", "eye_stability", "beard_edge_stability", "hand_integrity", "clothing_stability", "geometry_stability", "subject_background_integration", "zero_shimmer", "zero_warp", "stable_object_contact", "stable_face_outline", "stable_neck_and_collar", "stable_material_reflections" ], "forbidden": [ "latent_AI_skin", "synthetic_eyes", "overpolished_face", "ambiguous_facial_texture", "unstable_hands", "unclear_edges", "flicker_prone_glossy_face", "AI_edge_halos", "clothing_seams_likely_to_crawl", "glossy_zones_likely_to_flicker" ] }, "validation_gate": { "must_pass": [ "same_person_as_IMAGE1", "same_skin_truth_as_IMAGE1", "same_body_reality_as_IMAGE2", "correct_IMAGE3_main_subject_classification_for_transfer_only", "correct_conditional_IMAGE3_transfer", "if_IMAGE3_is_male_coherent_wardrobe_clothing_accessories_and_objects_are_copied_when_physically_plausible", "if_IMAGE3_is_female_clothing_wardrobe_accessories_jewelry_makeup_hairstyle_and_worn_carried_attached_objects_are_blocked", "if_IMAGE3_has_no_human_or_is_uncertain_IMAGE2_wardrobe_body_are_preserved_while_scene_camera_lights_are_copied", "no_AI_skin", "no_plastic_skin", "no_waxy_skin", "no_makeup_invention", "no_generic_handsome_face", "no_beauty_retouch", "credible_real_human_perception", "Canon_EOS_R50_real_camera_feeling", "usable_as_master_frame_for_Kling3" ], "reject_if": [ "looks_beautified", "looks_cleaner_than_IMAGE1", "skin_tone_more_uniform_than_IMAGE1", "beard_looks_cleaner_than_IMAGE1", "looks_like_different_person", "feels_synthetic", "style_from_IMAGE3_contaminates_skin", "male_IMAGE3_clothing_not_copied_when_coherent", "female_IMAGE3_wardrobe_or_accessories_copied", "female_IMAGE3_carried_worn_handheld_attached_or_body_dependent_objects_copied", "accessory_breaks_hands_or_identity", "object_unstable_for_Kling", "scene_looks_CGI", "lighting_looks_poster_render", "pose_deforms_body", "wardrobe_changes_body_mass", "unstable_for_video" ] }, "render_priority": [ "FACE_IDENTITY_FROM_IMAGE1", "BODY_ANATOMY_FROM_IMAGE2", "IMAGE3_MAIN_SUBJECT_CLASSIFICATION_FOR_TRANSFER_ONLY", "IMAGE3_TRANSFER_DECISION_MATRIX", "CONDITIONAL_IMAGE3_TRANSFER_BY_SUBJECT_TYPE", "REAL_REFERENCE_STRUCTURE_PRIORITY", "SKIN_TRUTH_MICROTEXTURE_FROM_IMAGE1", "FINAL_HUMAN_PERCEPTION", "MALE_IMAGE3_WARDROBE_ACCESSORIES_OBJECTS_WHEN_COHERENT", "FEMALE_IMAGE3_SCENE_POSE_LIGHT_STYLE_ONLY_WITH_WARDROBE_ACCESSORIES_OBJECTS_BLOCKED", "NO_HUMAN_OR_UNCERTAIN_IMAGE3_SCENE_CAMERA_LIGHT_ONLY", "ANTI_POSE_BODY_DEFORMATION", "STYLE_FIREWALL", "ANTI_GENERATED_SCENE_FINISH", "CANON_EOS_R50_REALISM", "KLING3_STABILITY" ], "final_rule": "Reject immediately if the result looks beautified, polished, synthetic, glamorous, too clean, too smooth, actor-like, model-like, luxury-ad-like, mannequin-like, CGI-like, plastic, waxy, oily, makeup-covered, showroom-rendered, poster-like, gender-contaminated, anatomically forced, falsely hyperreal, or AI-generated. Accept only if IMAGE1 identity and micro-skin truth remain intact, IMAGE2 body anatomy and scale remain intact, IMAGE3 transfer follows main-subject classification and decision-matrix logic, male IMAGE3 wardrobe/accessories/objects transfer when coherent, female IMAGE3 wardrobe/accessories/worn-carried-attached objects are fully blocked, real reference structure is preserved, scene avoids generated perfection, and final output feels like a real imperfect human captured by a Canon EOS R50 and stable enough for Kling 3.0." }
1139
Nano Banana
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
507522
Nano Banana
gett nude Learn to create better images with AI! Check out our crash course in prompt engineering & AI art generation! AM1970 posted about 2 hours ago 8 views 0 comments Behold this realistic masterpiece: a slim, elegant woman with fiery red hair tied in a long ponytail, defined eyebrows framing her striking blue eyes, and pale skin dotted with charming freckles across her face and chest. Captured from a side angle, she radiates happiness with a closed-mouth smile, her upturned nose adding to her allure. Dressed in a vibrant, colorful fabric robe pieced together with zips, buttons, seams, and intricate stitching—fully clothed yet teasing partial nudity with subtle cleavage and teardrop-shaped, saggy breasts for a natural, confident vibe. Illuminated by dramatic rim lighting that highlights every textured, detailed inch of her perfect skin, this high-quality image is a celebration of beauty, realism, and artistry! Read more... Comments (0) Add a comment Stunning handsome readhead woman by AM1970 • Follow 0 realistic, masterpiece, high quality BREAK 1woman, 28 y. o. slim fit woman, red hair, long pony tail, defined eyebrows, (fully clothed) , pale skin, (realistic skin:1.3) | (textured skin:1.3) | (detailed skin:1.3) | (perfect skin:1.3) , (upturned nose) , freckles on face and chest, colorful fabric combining to form robe zips, buttons, seams, stitching), partial nudity, teardrop boobs, blue eyes, (cleavage:1.2) , saggy breasts. side angle. happy| closed mouth. ((rim lighting) ) Negative prompt: unrealistic, saturated, high contrast, big nose, painting, drawing, sketch, cartoon, anime, manga, render, CG, 3d, watermark, signature, label, nsfw, nudity, fetish, latex, pvc, suggestive pose, weapon, combat, blood, war scene, aggressive expression, harsh shadows, heavy makeup, color tint, saturation, chromatic aberration, lens flare, low resolution, blur, noise, artifacts, over-retouching, plastic skin, oversharpening, watermark, text, logo Generation parameters 832x1216 Model used CyberRealistic XL v7.0 [Download AI model] More by AM1970 Twitter Instagram Say hi 👋 PromptHero is built by ElevenYellow . Navigation Create AI images Academy - Learn AI PromptHero Community Affiliate program Openjourney AI Challenges Verified profiles Plans Advertise FAQ 🟢 Status Learn AI Content Creation with AI for Brands and Products Midjourney for graphic design & art professionals Crash course in generative AI & prompt engineering for images AI influencers and consistent characters Create custom AI models and LoRas by fine-tuning Stable Diffusion Master your composition: advanced AI image generation with ControlNet Guides Best AI Upscalers 2025 Best AI Video Editors 2025 Best AI 3D Models 2025 Best AI Image Editors 2025 Best AI Notetakers 2025 The Best AI Prompts Best AI Prompts Best Veo Prompts Best Hailuo Prompts Best Wanx Prompts Best Hunyuan Prompts Best Seedance Prompts Best Pixverse Prompts Best Luma Prompts Best Kling Prompts Best Runway Prompts Best Nano Banana Prompts Best SeedEdit Prompts Best Qwen Prompts Best FLUX Prompts Best Recraft Prompts Best Ideogram Prompts Best Seedream 4 Prompts Best Google Imagen Prompts Best ChatGPT Image Prompts Best Stable Diffusion Prompts Best DreamShaper Prompts Best Realistic Vision Prompts Best epiCRealism Prompts Best majicMIX realistic Prompts Best Midjourney Prompts Best Openjourney Prompts Best ChilloutMix Prompts Best Deliberate Prompts Best ReV Animated Prompts Best Perfect World Prompts Best Dreamlike Photoreal Prompts Best Dreamlike Diffusion Prompts Best Anything Prompts Best Protogen Prompts Best AbyssOrangeMix Prompts Best Grapefruit Prompts Best Kenshi Prompts Best Analog Diffusion Prompts Best SynthwavePunk Prompts Best Illuminati Diffusion Prompts Best Arcane Diffusion Prompts Best Mo-Di Diffusion Prompts Best Redshift Diffusion Prompts Best Waifu Diffusion Prompts Best TrinArt Prompts Best Poolsuite Diffusion Prompts Best Funko Diffusion Prompts Best Text to Pokémon Prompts Best DALL-E Prompts Best Fashion Diffusion Prompts AI Models All AI models Image generation AI Models Video generation AI Models Large Language Models LoRA Textual Inversion ControlNet Hypernetwork Aesthetic Gradient LyCORIS VAE Poses Wildcards Workflows Other Other sites Dreamshot Jobician 2025 © PromptHero @prompthero
1362
FLUX
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
6516
Nano Banana
{ "prompt_type": "descriptive_replication", "reference_adherence": "STRICT_VISUAL_FIDELITY", "aspect_ratio": "3:4", "identity_lock": { "priority": "NONE", "instruction": "New female model. Completely different facial structure and overall appearance from any previous version. No resemblance to prior identity." }, "scene_constraints": { "people_count": "exactly one person in the frame", "single_subject_only": true, "no_other_people_visible": true, "no_extra_reflections": true, "no_background_figures": true, "full_body_visible": true, "head_to_toe_in_frame": true }, "subject": { "demographics": "young woman (age 20+), blonde", "presence": "only one visible person in the entire scene", "hair": { "color": "warm blonde (no dark roots)", "style": "side part, long straight hair", "texture": "smooth, natural fall, lightly voluminous" }, "face": { "structure": "distinct oval face shape with higher cheekbones and a slightly sharper jawline", "expression": "calm alluring expression with a subtle closed-lip smirk", "eyes": "almond-shaped eyes with a slightly lifted outer corner", "gaze": "looking at the phone screen through the mirror with composed confidence", "nose": "straight narrow nose with refined bridge", "lips": "naturally fuller lips with soft definition", "makeup": "minimal natural makeup, soft matte skin finish, muted rose lips, softly blended eyeshadow", "visibility": "partially covered by phone but clearly a different facial identity than any previous version" } }, "body": { "pose": "standing in front of mirror with one arm raised above head, legs naturally extended", "posture": "relaxed confident stance with slight hip shift", "anatomy": { "curves": "natural soft curves with realistic proportions", "chest": "bikini-supported, natural shape", "details": "visible collarbones, subtle shoulder definition, natural leg proportions" }, "feet": { "visibility": "fully visible at bottom of frame", "position": "flat on floor, natural stance" }, "skin_texture": "natural skin texture, no airbrushing" } }, "attire_and_accessories": { "clothing": { "top": "black triangle string bikini top with thin straps", "bottom_detail": "matching bikini bottom strings visible at hips" }, "footwear": "barefoot", "jewelry": { "necklaces": [], "wrists": [], "fingers": "no rings" }, "props": { "phone": "white iPhone (iPhone Pro style), held in right hand" } }, "environment": { "setting": "modern luxury bathroom interior, daytime", "foreground_element": { "mirror": "large rectangular mirror with GREY MARBLE FRAME", "material": "polished stone marble with subtle veining" }, "background": { "elements": [ "large marble wall tiles", "floating vanity", "minimalist sink and chrome fixtures", "folded white towels", "clean tiled floor fully visible" ], "style": "minimalist and clean", "clutter": "none" } }, "lighting_and_atmosphere": { "source": "natural daylight from camera-left", "quality": "soft directional light", "effects": [ "subtle highlights on shoulder and collarbone", "gentle shadow falloff along legs and feet" ], "contrast": "moderate, natural" }, "camera_and_technical": { "perspective": "mirror selfie", "camera_position": "eye level, slightly angled downward", "framing": "vertical 3:4, full head-to-toe body visible without cropping feet", "focus": "sharp on subject, background slightly softer", "visual_fidelity": "real smartphone camera look, mild natural noise" }, "realism_constraints": { "allowed": [ "natural skin texture", "minor asymmetry", "realistic lighting" ], "forbidden": [ "multiple people", "extra person", "background people", "beauty filters", "airbrushed skin", "anime style", "cartoon style", "cropped feet", "cut-off legs" ] } }
1356
Nano Banana
Ultra-realistic 8K scene of the person in the photo, stepping out of and breaking the panel border of a giant open manga book lying on the floor of an aesthetic bedroom. The comic pages depict a surreal 4-panel black-and-white manga with screentones, bold ink lines, and authentic Japanese layout, featuring the manga version of the same person trapped inside. The larger, real-life version emerges dynamically into the comic’s surface, vibrant, colorful, and photorealistic, in sharp contrast to the monochrome manga world. Both versions interact directly, creating a meta, humorous encounter. Highly detailed, crisp manga line art seamlessly integrated with vivid, realistic
24964
Nano Banana
Learn to create better images with AI! Check out our crash course in prompt engineering & AI art generation! AM1970 posted about 2 hours ago 8 views 0 comments Behold this realistic masterpiece: a slim, elegant woman with fiery red hair tied in a long ponytail, defined eyebrows framing her striking blue eyes, and pale skin dotted with charming freckles across her face and chest. Captured from a side angle, she radiates happiness with a closed-mouth smile, her upturned nose adding to her allure. Dressed in a vibrant, colorful fabric robe pieced together with zips, buttons, seams, and intricate stitching—fully clothed yet teasing partial nudity with subtle cleavage and teardrop-shaped, saggy breasts for a natural, confident vibe. Illuminated by dramatic rim lighting that highlights every textured, detailed inch of her perfect skin, this high-quality image is a celebration of beauty, realism, and artistry! Read more... Comments (0) Add a comment Stunning handsome readhead woman by AM1970 • Follow 0 realistic, masterpiece, high quality BREAK 1woman, 28 y. o. slim fit woman, red hair, long pony tail, defined eyebrows, (fully clothed) , pale skin, (realistic skin:1.3) | (textured skin:1.3) | (detailed skin:1.3) | (perfect skin:1.3) , (upturned nose) , freckles on face and chest, colorful fabric combining to form robe zips, buttons, seams, stitching), partial nudity, teardrop boobs, blue eyes, (cleavage:1.2) , saggy breasts. side angle. happy| closed mouth. ((rim lighting) ) Negative prompt: unrealistic, saturated, high contrast, big nose, painting, drawing, sketch, cartoon, anime, manga, render, CG, 3d, watermark, signature, label, nsfw, nudity, fetish, latex, pvc, suggestive pose, weapon, combat, blood, war scene, aggressive expression, harsh shadows, heavy makeup, color tint, saturation, chromatic aberration, lens flare, low resolution, blur, noise, artifacts, over-retouching, plastic skin, oversharpening, watermark, text, logo Generation parameters 832x1216 Model used CyberRealistic XL v7.0 [Download AI model] More by AM1970 Twitter Instagram Say hi 👋 PromptHero is built by ElevenYellow . Navigation Create AI images Academy - Learn AI PromptHero Community Affiliate program Openjourney AI Challenges Verified profiles Plans Advertise FAQ 🟢 Status Learn AI Content Creation with AI for Brands and Products Midjourney for graphic design & art professionals Crash course in generative AI & prompt engineering for images AI influencers and consistent characters Create custom AI models and LoRas by fine-tuning Stable Diffusion Master your composition: advanced AI image generation with ControlNet Guides Best AI Upscalers 2025 Best AI Video Editors 2025 Best AI 3D Models 2025 Best AI Image Editors 2025 Best AI Notetakers 2025 The Best AI Prompts Best AI Prompts Best Veo Prompts Best Hailuo Prompts Best Wanx Prompts Best Hunyuan Prompts Best Seedance Prompts Best Pixverse Prompts Best Luma Prompts Best Kling Prompts Best Runway Prompts Best Nano Banana Prompts Best SeedEdit Prompts Best Qwen Prompts Best FLUX Prompts Best Recraft Prompts Best Ideogram Prompts Best Seedream 4 Prompts Best Google Imagen Prompts Best ChatGPT Image Prompts Best Stable Diffusion Prompts Best DreamShaper Prompts Best Realistic Vision Prompts Best epiCRealism Prompts Best majicMIX realistic Prompts Best Midjourney Prompts Best Openjourney Prompts Best ChilloutMix Prompts Best Deliberate Prompts Best ReV Animated Prompts Best Perfect World Prompts Best Dreamlike Photoreal Prompts Best Dreamlike Diffusion Prompts Best Anything Prompts Best Protogen Prompts Best AbyssOrangeMix Prompts Best Grapefruit Prompts Best Kenshi Prompts Best Analog Diffusion Prompts Best SynthwavePunk Prompts Best Illuminati Diffusion Prompts Best Arcane Diffusion Prompts Best Mo-Di Diffusion Prompts Best Redshift Diffusion Prompts Best Waifu Diffusion Prompts Best TrinArt Prompts Best Poolsuite Diffusion Prompts Best Funko Diffusion Prompts Best Text to Pokémon Prompts Best DALL-E Prompts Best Fashion Diffusion Prompts AI Models All AI models Image generation AI Models Video generation AI Models Large Language Models LoRA Textual Inversion ControlNet Hypernetwork Aesthetic Gradient LyCORIS VAE Poses Wildcards Workflows Other Other sites Dreamshot Jobician 2025 © PromptHero @prompthero
1226
FLUX
<think> The user wants a Z-Image-Turbo prompt based on the provided JSON description. I need to extract information from each field in the JSON and structure it into a prompt following the specified format. Let's break down each section and map it to the prompt structure: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - From "description": stunning model in her early 20s, European features with fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on arms and cheeks, natural micro-imperfections, light natural freckles across nose and upper chest, lightly tanned sun-kissed glow on collarbones and shoulders, athletic yet feminine body with realistic natural proportions, medium perky breasts with soft realistic shape and subtle natural cleavage visible through slightly sheer sections of her form-fitting white and black tactical outfit, low-cut neckline and tight fit accentuating toned midriff, slim waist, defined hips and long legs. - From "face": Long tousled blonde hair with natural volume, sun-kissed highlights and individual strands moving gently as she walks, extremely detailed large beautiful brilliant blue-silver-golden-grey eyes with intricate iris textures, realistic reflections and natural depth, eyes fluttering while staring directly into the camera with intense captivating gaze, full glossy lips transitioning into a warm radiant genuine smile over the scene showing perfect white teeth, high cheekbones, soft natural blush on cheeks. - From "outfit": Form-fitting white and black tactical outfit with futuristic design: high-tech fabric with subtle matte-gloss contrast, deep V-neckline and strategic sheer mesh panels revealing skin tone and natural curves underneath for subtle NSFW sensuality, tactical straps and utility details on shoulders and hips, black accents on sleeves and waist, no visible underwear lines for clean realistic look. 2. **Action or scene context** - From "pose": Dynamic confident walk down a futuristic corridor, hips swinging gracefully, one leg forward in mid-stride, arms relaxed yet purposeful, head slightly tilted toward camera with engaging direct eye contact and growing smile. 3. **Environment / setting + time of day + weather** - From "environment": Sleek futuristic corridor with metallic walls reflecting soft light, continuous yellow glowing floor line creating strong perspective and depth, vertical warm amber light panels on walls casting cinematic illumination, subtle sci-fi ambient fog in distance, high-tech yet elegant atmosphere. 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - From "lighting": Warm diffused lighting from vertical panels on both sides, soft golden-amber glow blending with cool metallic reflections, gentle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing three-dimensional form and skin texture, high dynamic range with vivid yet natural colors. 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - From "camera": Cinematic full-body shot with medium telephoto lens (85–135mm equivalent), f/2.0 aperture, razor-sharp focus from face to feet, shallow depth of field with creamy smooth bokeh on corridor background, slight low angle to emphasize power and grace, perfect viewpoint. 6. **Mood / atmosphere / emotion** - From "style" and "environment": elegant futuristic atmosphere with graceful and seductive vibe. 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - From "style": Ultra photorealistic cinematic photography. - From "quality": trending professional sci-fi fashion portrait. 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - From "quality": 8K resolution, extreme sharpness, rich cinematic colors with warm amber tones and cool metallic contrast, highly detailed skin realism, hair movement, outfit texture, corridor reflections and light play. - From "style" and general understanding: masterpiece quality. Now, assemble the prompt in the specified format. </think> A stunning young woman with European features and fair smooth skin showcasing hyper-detailed realistic textures, including visible fine pores, subtle peach fuzz, natural micro-imperfections, light freckles, sun-kissed glow on collarbones, athletic yet feminine body with realistic proportions, medium perky breasts with subtle natural cleavage visible through sheer sections of a form-fitting white and black tactical outfit, low-cut neckline accentuating toned midriff, slim waist, defined hips, long legs, long tousled blonde hair with sun-kissed highlights and individual strands gently moving, extremely detailed large brilliant blue-silver-golden-grey eyes with intricate iris textures and realistic reflections, full glossy lips transitioning into a warm radiant smile showing perfect white teeth, high cheekbones, soft natural blush on cheeks, dynamic confident walk down a futuristic corridor with hips swinging gracefully, one leg forward in mid-stride, arms relaxed yet purposeful, head tilted toward camera with engaging direct eye contact and growing smile, sleek futuristic corridor with metallic walls reflecting soft light, continuous yellow glowing floor line creating perspective and depth, vertical warm amber light panels on walls casting cinematic illumination, subtle sci-fi ambient fog in distance, high-tech yet elegant atmosphere, warm diffused lighting from vertical panels, soft golden-amber glow blending with cool metallic reflections, gentle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing form and skin texture, high dynamic range with vivid natural colors, cinematic full-body shot with medium telephoto lens (85–135mm equivalent), f/2.0 aperture, razor-sharp focus from face to feet, shallow depth of field with creamy smooth bokeh on corridor background, slight low angle emphasizing power and grace, perfect viewpoint, ultra photorealistic cinematic photography, maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture, realistic hair strands with natural movement, authentic fabric weave and subtle sheer effect, elegant futuristic atmosphere with graceful and seductive vibe, 8K resolution, extreme sharpness, rich cinematic colors with warm amber tones and cool metallic contrast, highly detailed skin realism, hair movement, outfit texture, corridor reflections and light play, trending professional sci-fi fashion portrait, masterpiece quality, ultra detailed, absurdres, intricate, sharp focus, depth of field, best quality, 8k uhd, hdr, photorealistic, realistic textures, dramatic rim lighting, volumetric fog, golden hour lighting, The subject is sharply depicted with outstanding detail., flux_pro
201525
Z Image
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
3246
Nano Banana
High-definition quality, photo version. Shot with a Canon EOS 5D Mark IV and an EF 100mm f/2.8L Macro IS USM lens, f/4.5, 1/200s, ISO 400. Best quality, ultra HD, 8K, award-winning photography. --ar 3:4 High-definition, realistic skin textures, extremely detailed facial textures, cinematic lighting, and a natural atmosphere, no cartoon style.
10353
Sora
Steps: 20, Sampler: DPM++ 2M Karras, CFG scale: 7, Seed: 413127992, Size: 1280x1600, Model hash: 7fa552ffb0, Model: nijidiffusedmix_v40, Denoising strength: 0.3, Clip skip: 2, ADetailer model: face_yolov8s.pt, ADetailer prompt: "dark blue eyes, freckles below eyes", ADetailer confidence: 0.3, ADetailer dilate erode: 4, ADetailer mask blur: 4, ADetailer denoising strength: 0.5, ADetailer inpaint only masked: True, ADetailer inpaint padding: 32, ADetailer version: 23.12.0, Version: v1.6.1
10493
Stable Diffusion
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
7821
Nano Banana
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
5393
Nano Banana
--- name: pastilles-llm-multi-agent description: Variante Claude Code du générateur de pastilles LLM, qui lance réellement plusieurs sous-agents en parallèle. Génère une pastille de communication interne sur les LLM (un texte court en français plus un prompt unique de génération d'images à coller dans le chat Gemini), via cinq sous-agents indépendants, une fusion, puis une revue critique par trois sous-agents et une correction. Utilise ce skill dès qu'on te demande de rédiger, produire ou générer une pastille, une fiche ou un contenu court d'acculturation sur les LLM, l'IA générative, le prompting, les agents, le RAG, la confidentialité IA ou tout sujet de la liste des 45 pastilles ci-dessous, que le mot "pastille" soit employé ou non. Utilise-le aussi dès qu'on te fournit un titre issu de cette liste. --- # Générateur de pastilles LLM, version multi-agents (Claude Code) ## Ce que fait ce skill Produit une pastille pédagogique complète à partir d'un titre. À la différence de la version chat, il lance de vrais sous-agents parallèles: une passe de recherche, puis cinq sous-agents indépendants qui rédigent chacun un brouillon sous un angle différent, puis une fusion par l'orchestrateur qui retient les meilleures formulations, et enfin une revue critique par trois sous-agents suivie d'une correction. Livrable: le texte de la pastille et un prompt unique de génération d'images à coller dans le chat Gemini. ## Environnement requis Ce skill suppose Claude Code, avec les sous-agents (outil Task) et la recherche web disponibles. Si les sous-agents ne sont pas disponibles dans ton environnement, n'improvise pas: utilise la variante chat, qui fait le même travail en self-ensemble séquentiel. ## Entrée Un titre de pastille, idéalement issu de la liste ci-dessous. Si le titre est ambigu ou hors liste, demande une clarification avant de générer. ## Liste des 45 pastilles (pour la continuité) 1. Au fait, c'est quoi un LLM ? 2. Modèles de fondation : ces géants pré-entraînés dont tout découle 3. Dans les coulisses : comment une IA apprend (sans vraiment comprendre) 4. Le contexte : la "mémoire vive" de l'IA (ce qu'elle voit à un instant T) 5. Pourquoi l'IA oublie : la vérité sur la mémoire d'une session à l'autre 6. Température et créativité : régler le curseur de la "folie" de l'IA 7. Les tokens : la monnaie d'échange (et la manière de penser) des LLM 8. Le non-déterminisme : pourquoi l'IA fait des stats et ne donne jamais deux fois la même réponse 9. Multimodalité : quand l'IA commence à voir, entendre et coder 10. Knowledge cutoff : pourquoi votre IA vit (parfois) dans le passé 11. Les biais de l'IA : le miroir pas toujours glorieux de nos propres données 12. Zéro pointé : pourquoi l'IA est structurellement nulle en calcul mental 13. Les hallucinations : quand l'IA invente la vie avec un aplomb incroyable 14. L'art du Fact-Checking : comment ne pas gober les hallucinations de l'IA 15. L'empreinte carbone de l'IA : combien d'eau et de CO2 pour un prompt ? 16. Anatomie d'un bon prompt : la recette de base 17. Les pièges du prompt : flou artistique et overdose d'instructions 18. L'art de la discussion : pourquoi il faut itérer plutôt que tout demander d'un coup 19. Donner des ordres : comment imposer un format de sortie strict (JSON, tableaux, etc.) 20. Diviser pour régner : découper une tâche complexe en étapes claires 21. Le Few-Shot Prompting : un bon exemple vaut mille consignes 22. Inception : donner un rôle et un contexte à votre IA 23. Zéro syndrome de la page blanche : rédiger et reformuler ses mails ou CR 24. La diagonale du fou : résumer et synthétiser un pavé de 50 pages en 2 minutes 25. Le sparring-partner : brainstormer et structurer ses idées sans filtre 26. Caméléon : traduire et adapter son ton selon le destinataire 27. Data-analyst de poche : faire parler un tableur ou des données brutes 28. Mise en situation : préparer une réunion ou un brief client avec l'IA 29. Avis aux dévs : l'IA comme copilote de code et de review 30. SLM : Pourquoi sortir un tank pour écraser une mouche ? 31. Le "harnais" : donner des bras, des yeux et un cadre au LLM 32. C'est quoi un "Agent" IA, concrètement ? 33. Human-in-the-loop : pourquoi il faut toujours garder un humain aux manettes 34. Quand l'agent s'emballe : autonomie et erreurs en cascade 35. Automatisation : confier un workflow complet à un agent 36. Tool Use : quand l'IA apprend à utiliser une calculatrice ou une API 37. Le RAG : faire réviser vos propres documents à l'IA avant qu'elle ne réponde 38. MCP et connecteurs : brancher l'IA directement sur vos outils de dev et d'orga 39. Responsabilité : vous êtes le seul signataire de ce que produit l'IA 40. Confidentialité : où vont vraiment les données que vous tapez ? 41. Alerte Rouge : les données clients à ne JAMAIS copier-coller dans une IA public 42. Shadow AI : pourquoi le ChatGPT gratuit perso au bureau est une fausse bonne idée 43. RGPD et IA : ce qu'il faut savoir pour rester dans les clous 44. Prompt Injection : quand on pirate une IA avec de simples phrases 45. La Chaîne de Pensée : pourquoi l'IA a besoin de "réfléchir à voix haute" Sers-toi de cette liste pour situer la pastille et délimiter son périmètre. Avant de rédiger: - Repère les 1 à 3 pastilles voisines qui recouvrent le sujet, et dresse une courte liste "déjà traité ailleurs, à ne pas ré-expliquer". - Si une pastille voisine a déjà été produite, demande son texte à l'utilisateur (ou appuie-toi dessus s'il est fourni) pour caler précisément la limite. - Les concepts fondateurs déjà couverts ne sont rappelés qu'en une phrase de mise en contexte, jamais ré-expliqués. Le coeur de la pastille est consacré à ce que son titre ajoute de neuf. - Ne renvoie pas vers les autres pastilles dans le texte final, sauf si le titre l'impose. ## Règles du texte - Longueur adaptée à la profondeur du sujet. Un sujet léger tient en 3 paragraphes courts; un sujet plus riche peut aller jusqu'à 4 paragraphes plus étoffés. Ne gonfle pas artificiellement un sujet simple et ne compresse pas à l'excès un sujet dense: juge la profondeur par la richesse réelle du concept. Prose uniquement, pas de listes ni de puces. - Ton décontracté, précis et léger: accessible, vivant et sans lourdeur, mais techniquement juste. Public visé: une main-d'oeuvre diverse en société de conseil, du profil non technique au développeur. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Reste léger, évite le name-dropping. N'accumule pas les noms de modèles, d'outils, d'entreprises, de chercheurs ou de techniques pour faire savant: privilégie l'explication claire du mécanisme. Ne cite un nom précis que s'il éclaire vraiment le propos, une ou deux fois au maximum. La recherche sert l'exactitude et alimente la section Sources, elle n'a pas à truffer le texte de références. - Contenu autonome: la pastille se comprend seule, sans avoir lu les autres, sauf mention contraire dans le titre. - Rédige en français. - Ne mentionne jamais ANEO ni aucun nom d'entreprise. - Le texte explicatif ne figure jamais dans l'image. - N'utilise pas de tiret cadratin. Privilégie des caractères standard (virgules, deux-points, parenthèses). ## Processus multi-agents ### Étape 1, recherche (orchestrateur, une seule fois) Toi, l'orchestrateur, fais d'abord une passe de recherche web ciblée sur le sujet de la pastille. Une seule passe, en amont. Les sous-agents ne peuvent pas lancer de sous-agents et n'héritent pas de ton contexte, donc c'est à toi de faire la recherche puis de leur transmettre les résultats. - Objectif: exactitude, chiffres, exemples concrets et faits à jour. - Ancre tes recherches dans le présent: repère la date du jour fournie dans ton contexte (champ currentDate) et privilégie les informations et chiffres les plus récents, car ce domaine évolue vite. Précise cette date au besoin dans tes requêtes. - Priorise les sources originales ou officielles (blogs d'éditeurs, articles de recherche, sites gouvernementaux, documentation) plutôt que les agrégateurs. - Adapte la profondeur: quelques recherches pour un concept stable (par exemple ce qu'est un token), davantage pour un sujet à chiffres ou mouvant (empreinte carbone et coûts, modèles spécifiques, SLM, MCP et connecteurs, RGPD). - Produis un brief de recherche compact: faits clés, chiffres utiles, et 2 à 4 sources. Tu l'injecteras tel quel dans chaque sous-agent, et tu listeras les sources dans le livrable final. ### Étape 2, affinement de la thématique (orchestrateur, une seule fois) Si plusieurs axes apparaissent pertinents après avoir enlever les axes adressés par d'autres pastilles prévues, dialogue avec l'utilisateur pour choisir l'axe de la pastille. Si l'axe choisi nécessite de nouvelles sources, refait en une recherche web selon les consignes précédentes. ### Étape 3, fan-out (cinq sous-agents en parallèle) Lance cinq sous-agents via l'outil Task, dans le même tour, pour qu'ils s'exécutent en parallèle. Sois explicite sur le parallélisme: par défaut Claude Code reste séquentiel, il ne parallélise que si tu le demandes clairement. Un sous-agent par angle: 1. Analogie: expliquer le concept via une métaphore concrète du quotidien. 2. Cas d'usage: partir d'une situation de travail réelle et vécue. 3. Idée reçue: démarrer d'un malentendu courant, puis rectifier. 4. Mécanique: aller au coeur du fonctionnement, précis mais accessible. 5. Enjeu: pourquoi ça compte, ce que ça change concrètement pour l'utilisateur. Comme le contexte n'est pas hérité, chaque prompt de sous-agent doit être autonome et contenir tout le nécessaire. Utilise ce gabarit, en remplaçant les crochets: ``` Tu rédiges un brouillon d'une pastille pédagogique interne sur les LLM. Tu travailles seul, sans accès au reste de la conversation. Titre de la pastille: [TITRE] Angle imposé: [ANGLE, par exemple "Analogie: expliquer via une métaphore concrète du quotidien"] À NE PAS ré-expliquer (déjà couvert par d'autres pastilles), à mentionner en une phrase au maximum: [LISTE des points déjà traités ailleurs, à ne pas retraiter] Brief de recherche (appuie-toi dessus, ne relance pas de recherche): [BRIEF: faits clés, chiffres, sources] Règles: - Longueur selon la profondeur du sujet: 3 paragraphes courts si le sujet est léger, jusqu'à 4 paragraphes étoffés s'il est dense. Prose uniquement, pas de listes. - Mets en gras et/ou en italique les termes importants, de l'ordre de 3 à 4 emphases par paragraphe (repère indicatif, ne force pas ce compte). - Ton décontracté, précis et léger, accessible mais techniquement juste. Évite le name-dropping: pas d'accumulation de noms de modèles, outils, entreprises ou chercheurs, privilégie l'explication du mécanisme. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Français. Ne mentionne aucune entreprise. Pas de tiret cadratin. - Contenu autonome, compréhensible seul. Réponds exactement dans ce format: TEXTE: [les 3 à 4 paragraphes] ILLUSTRATION_TITRE: [une phrase décrivant le concept central à illustrer] SCHEMA: [oui ou non. Si oui, décris en 2 à 3 lignes le diagramme et ses libellés en français] ``` ### Étape 4, fan-in et fusion (orchestrateur) Attends les cinq retours, puis fusionne en une seule pastille finale: - Retiens la structure la plus claire, la meilleure analogie, l'exemple le plus parlant, la correction la plus nette, la formulation la plus précise et le "pourquoi ça compte" le plus fort. - Réécris en une seule voix cohérente. Pas d'effet patchwork. - Respecte la longueur adaptée à la profondeur (voir Règles du texte) et le ton décontracté, précis et léger, sans name-dropping. - Décide s'il faut un schéma: inclus-en un si au moins deux sous-agents ont répondu "oui", ou si le concept est intrinsèquement un processus, un flux ou une comparaison (par exemple RAG, agents, étapes de prompting, comparaison de modèles). Le cas échéant, fusionne les meilleures idées de schéma. - Construis ensuite le prompt image unique (section ci-dessous). ### Règles d'écriture pour la pastille finale - Longueur selon la profondeur du sujet: 3 paragraphes courts si le sujet est léger, jusqu'à 4 paragraphes étoffés s'il est dense. Prose uniquement, pas de listes. - Mets en gras et/ou en italique les termes importants, de l'ordre de 3 à 4 emphases par paragraphe (repère indicatif, ne force pas ce compte). - Ton décontracté, précis et léger, accessible mais techniquement juste. Évite le name-dropping: pas d'accumulation de noms de modèles, outils, entreprises ou chercheurs, privilégie l'explication du mécanisme. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Français. Ne mentionne aucune entreprise. Pas de tiret cadratin. - Contenu autonome, compréhensible seul. ### Étape 5, revue critique et correction (trois sous-agents en parallèle, puis orchestrateur) Quand la lancer: systématiquement à la première génération, une fois le texte fusionné et le prompt image construits. Ensuite, si l'utilisateur demande des ajustements, ne relance pas la revue d'office: propose-la, et ne la lance qu'avec son accord (elle coûte trois sous-agents de plus). Principe: les relecteurs produisent des CONSTATS, jamais une réécriture. Chacun renvoie une liste de problèmes localisés, assortis d'une gravité et d'un correctif précis. C'est toi, l'orchestrateur, qui appliques et réécris, pour garder une voix unique et éviter l'effet patchwork. Ce que la revue peut juger: les relecteurs ne voient pas les images, générées plus tard dans Gemini. La revue du visuel porte donc uniquement sur le PROMPT image (clarté, cohérence avec le texte, respect des consignes), jamais sur un rendu. Le contrôle visuel réel reste à l'utilisateur. Lance trois sous-agents en parallèle (parallélisme explicite, sinon l'exécution sera séquentielle), un par grille. Le contexte n'étant pas hérité, chaque prompt doit être autonome et contenir tout le nécessaire: le titre, le texte fusionné, le bloc prompt image, le brief de recherche, la liste "déjà traité ailleurs" (et les textes des pastilles voisines s'ils sont disponibles), et la liste des 45 titres pour repérer les chevauchements. Les trois grilles: 1. Fond, exactitude et périmètre: exactitude vs le brief (aucun chiffre, date ou fait inventé ni sur-affirmé, rien qui le contredise), cohérence entre le titre et ce que le texte délivre, chevauchements avec les pastilles voisines (le texte ré-explique-t-il ce qui est traité ailleurs ? redites à signaler), autonomie du contenu, clarté du message à retenir, repérage de ce qui vieillira mal (à nuancer). 2. Forme, ton et pédagogie: ton décontracté-précis-léger et techniquement juste, absence de name-dropping, rythme et fluidité à la lecture à voix haute, longueur adaptée à la profondeur (3 à 4 paragraphes, prose, pas de listes) et chasse au verbiage, accessibilité pour un profil non technique (jargon expliqué, analogies claires), présence d'emphases utiles (gras/italique), de l'ordre de 3 à 4 par paragraphe à titre indicatif (signale un texte trop peu emphasé autant qu'un excès, mais ne fais jamais retirer une emphase pertinente), force de l'accroche. 3. Conformité et visuel: contraintes dures (aucun tiret cadratin ni caractère non standard, aucun nom d'entreprise ni ANEO, français correct, prose sans listes ni puces); puis le prompt image: titre exact au caractère près, charte présente une seule fois, texte de la pastille bien marqué "à ne pas afficher", illustration-titre iconique et non schéma de processus, schéma en 2e image séparée si retenu, libellés de schéma courts (groupes nominaux) et en français, cohérence entre le visuel décrit et le coeur du texte. Gabarit de relecteur (remplace les crochets): ``` Tu relis un brouillon quasi final de pastille pédagogique interne sur les LLM. Tu travailles seul, sans accès au reste de la conversation. Tu ne réécris pas: tu rends des constats et des correctifs précis. Titre de la pastille: [TITRE] Texte fusionné à relire: [TEXTE COMPLET] Prompt image associé (à relire aussi): [BLOC PROMPT IMAGE] Brief de recherche de référence (fait foi pour l'exactitude): [BRIEF] Déjà traité par d'autres pastilles, à ne pas ré-expliquer (et textes voisins si fournis): [LISTE + TEXTES VOISINS] Liste des 45 titres (pour repérer les chevauchements): [LISTE DES 45 TITRES] Ta grille de relecture (tiens-t'en strictement à elle): [RUBRIQUE 1, 2 OU 3] Rappel: tu ne peux pas voir d'image; juge seulement le PROMPT image, pas un rendu. Réponds exactement dans ce format: CONSTATS: - [bloquant|recommandé|mineur] [problème localisé] -> [correctif précis] - ... (n'invente pas de problème: si un point est bon, ne le liste pas) VERDICT: [une phrase: publiable tel quel / corrections mineures / corrections nécessaires] ``` Fan-in et correction (orchestrateur): rassemble les constats, dédoublonne, arbitre les conseils contradictoires. Garde-fou anti-gonflement: entre "ajouter" et "raccourcir", la concision l'emporte, sauf erreur de fond avérée. Applique les constats bloquants et recommandés, écarte ou mentionne les mineurs. Une seule passe, pas de boucle. Réécris la version corrigée en une seule voix, puis prépare le court résumé "Ce que la revue a corrigé" (voir Format de sortie). ### Points de vigilance Claude Code - Parallélisme explicite: demande clairement cinq sous-agents en parallèle, sinon l'exécution sera séquentielle. - Contexte non hérité: mets tout dans le prompt du sous-agent, en particulier le brief de recherche. - Pas d'imbrication: un sous-agent ne peut pas en lancer un autre, garde recherche et fusion chez l'orchestrateur. - Coût et limites: cinq sous-agents en parallèle multiplient l'usage de jetons et peuvent déclencher des limites de débit. En cas de limite, replie-toi sur une exécution séquentielle. - Revue: la phase de revue ajoute trois sous-agents. Elle tourne d'office à la première génération; sur les demandes d'ajustement ultérieures, propose-la et ne la lance qu'avec l'accord de l'utilisateur. En cas de limite de débit, exécute les relecteurs en séquentiel ou replie-toi sur un relecteur global unique. - Réemploi optionnel: tu peux définir les cinq angles comme sous-agents personnalisés dans .claude/agents/ pour les réutiliser, mais ce n'est pas obligatoire, les sous-agents polyvalents suffisent. ## Prompt de génération d'images (un seul bloc, à coller dans le chat Gemini) Produis un seul bloc de prompt, rédigé en français, à coller tel quel dans le chat de Gemini. Gemini se charge d'appeler Nano Banana une ou deux fois, ce qui réduit les manipulations pour l'utilisateur. Le bloc demande: - Toujours une illustration-titre. - Un schéma seulement s'il a été retenu à l'étape de fusion, généré comme une seconde image séparée (fichier distinct), jamais dans la même image que le titre. Contenu du bloc: - Décris la charte graphique une seule fois (bloc ci-dessous) et précise que toutes les images la partagent, pour une cohérence visuelle. - Précise: tout texte affiché dans les images est en français. Cette consigne est importante, elle évite qu'un libellé se retrouve en anglais. - Inclus le texte complet de la pastille dans le bloc, en contexte de génération, clairement marqué comme à ne pas afficher. Il aide le modèle à comprendre le sujet et à choisir une illustration juste. Les seuls textes rendus visibles sont le titre exact et, pour un schéma, ses libellés. Le texte de la pastille ne doit jamais apparaître dans les images. - Illustration-titre: composition épurée et moderne, focus graphique central iconique représentant le concept. Ce n'est pas un schéma de processus. Le seul texte affiché est le titre exact, en en-tête: n'ajoute aucun sous-titre, accroche ou texte secondaire, le titre seul. Technique en deux temps pour le titre: dans le prompt généré, écris le titre exact entre guillemets droits (forme attendue: Le titre exact: "...") et demande un rendu fidèle, sans faute, en police sans serif corporate. Les guillemets ne sont qu'un délimiteur côté prompt: ne les commente pas et n'ajoute aucune consigne à leur sujet dans le prompt généré (pas de mention du type "les guillemets ne doivent pas apparaître"). - Schéma s'il est retenu: une seconde image générée séparément, dans son propre fichier, jamais fusionnée avec l'illustration-titre ni juxtaposée dans la même image. Diagramme d'entreprise propre et net (processus, flowchart ou comparaison), cohérent avec l'illustration-titre, mêmes charte et style, libellés fournis explicitement en français, très peu de fioritures. - Libellés de schéma: des groupes nominaux courts et lisibles seuls (par exemple "Exemples de bonnes réponses"), pas des phrases verbales qui sonnent comme des ordres ("Montrer de bonnes réponses"). Quelques mots par libellé. Attention aux titres longs: Nano Banana rend fiablement les libellés courts mais peut faire une faute sur un titre long; la technique en deux temps limite le risque. Si le rendu d'un titre long reste incertain, indique à l'utilisateur qu'il peut demander à Gemini d'utiliser Nano Banana Pro, plus fiable pour le texte long. Gabarit, illustration-titre seule (exemple avec le titre 1): ``` Génère une image en respectant cette charte graphique: [bloc Charte graphique]. Contexte pour comprendre le sujet, à NE PAS afficher dans l'image: [ici le texte complet de la pastille]. Illustration-titre pour une pastille pédagogique sur les LLM. Composition épurée et moderne, focus graphique central iconique: un nuage de lettres et de mots qui se condense en une bulle de dialogue, évoquant un modèle de langage qui transforme du texte en réponse. Ce n'est pas un schéma de processus. Seul texte à afficher dans l'image: le titre, en en-tête, sans faute d'orthographe, en police sans serif corporate, sans sous-titre ni texte secondaire. Le titre exact: "TITRE EXACT". Tout le texte affiché est en français. Format 16:9. ``` Gabarit, illustration-titre plus schéma (deux images): ``` Prépares-toi à générer deux images séparées, dans deux fichiers distincts. Les deux partagent cette charte graphique: [bloc Charte graphique]. Tout texte affiché dans les images est en français. Contexte pour comprendre le sujet, à NE PAS afficher dans les images: [ici le texte complet de la pastille]. Image 1, illustration-titre: composition épurée et moderne, focus graphique central iconique qui illustre le sujet. Ce n'est pas un schéma de processus. Seul texte à afficher: le titre, en en-tête, sans faute d'orthographe, police sans serif corporate, sans sous-titre ni texte secondaire. Le titre exact: "TITRE EXACT". Format 16:9. Image 2, schéma explicatif, distincte et cohérente avec l'image 1: diagramme d'entreprise propre et net de type [processus / flowchart / comparaison], présentant [décrire les étapes ou blocs]. Libellés exacts à afficher en français: [liste des libellés]. Très peu de fioritures, focus sur la clarté, n'inclus pas le titre. Génère dans un premier temps seulement la première image (l'illustration-titre), et attends les instructions de l'utilisateur pour générer la 2e image. ``` ## Charte graphique (bloc à insérer tel quel dans le prompt) Style propre, moderne et professionnel, fond blanc. Palette: orange vif et bleu corporate foncé en dominantes, blanc et gris pour les respirations, accents multiculturels discrets (rouge, vert, jaune, bleu) reprenant subtilement un motif de couleurs de logo. Superpositions graphiques épurées: motifs géométriques abstraits, quartiers de cercle, lignes claires. Composition aérée, jamais surchargée. Typographie sans serif, propre et corporate. ## Format de sortie Le travail des sous-agents reste dans leur propre contexte, seuls leurs retours te reviennent. N'affiche donc que le livrable final (déjà corrigé par l'étape de revue), la conversation principale reste propre. Si l'utilisateur demande à voir l'atelier, restitue les cinq brouillons reçus et les constats de revue. Livrable final, dans cet ordre: - Quand une revue a eu lieu (systématiquement à la première génération), un court résumé "Ce que la revue a corrigé" (2 à 4 lignes) présentant les principaux ajustements, avant le reste. Lors d'ajustements ultérieurs sans revue, présente simplement la version ajustée sans ce résumé. - Le texte de la pastille (3 à 4 paragraphes), dans sa version corrigée. - Un bloc de code intitulé "Prompt images (à coller dans Gemini)", contenant le prompt unique prêt à copier. - Une courte section "Sources" listant 2 à 4 références principales issues de l'étape de recherche, de préférence officielles ou originales. Cette section sert à la vérification et n'a pas vocation à être publiée dans la pastille. Termine toujours par cette question exacte: "Comment trouvez-vous le texte, l'image titre et le diagramme (si généré) ? Si une partie vous semble trop complexe, ou si vous souhaitez affiner le focus d'un visuel pour qu'il soit encore plus épuré, n'hésitez pas à me le faire savoir, et je l'ajusterai." -> le personnage doit incarner le professeur un peu foufou qui allusine et obsédé par l'ai
015
Grok Image
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
727622
Z Image
<think> The user wants a Z-Image-Turbo prompt based on the provided JSON description. I need to extract information from each field in the JSON and structure it into a prompt following the specified format. Let's break down each section and map it to the prompt structure: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - From "description": stunning model in her early 20s, European features with fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on arms and cheeks, natural micro-imperfections, light natural freckles across nose and upper chest, lightly tanned sun-kissed glow on collarbones and shoulders, athletic yet feminine body with realistic natural proportions, medium perky breasts with soft realistic shape and subtle natural cleavage visible through slightly sheer sections of her form-fitting white and black tactical outfit, low-cut neckline and tight fit accentuating toned midriff, slim waist, defined hips and long legs. - From "face": Long tousled blonde hair with natural volume, sun-kissed highlights and individual strands moving gently as she walks, extremely detailed large beautiful brilliant blue-silver-golden-grey eyes with intricate iris textures, realistic reflections and natural depth, eyes fluttering while staring directly into the camera with intense captivating gaze, full glossy lips transitioning into a warm radiant genuine smile over the scene showing perfect white teeth, high cheekbones, soft natural blush on cheeks. - From "outfit": Form-fitting white and black tactical outfit with futuristic design: high-tech fabric with subtle matte-gloss contrast, deep V-neckline and strategic sheer mesh panels revealing skin tone and natural curves underneath for subtle NSFW sensuality, tactical straps and utility details on shoulders and hips, black accents on sleeves and waist, no visible underwear lines for clean realistic look. 2. **Action or scene context** - From "pose": Dynamic confident walk down a futuristic corridor, hips swinging gracefully, one leg forward in mid-stride, arms relaxed yet purposeful, head slightly tilted toward camera with engaging direct eye contact and growing smile. 3. **Environment / setting + time of day + weather** - From "environment": Sleek futuristic corridor with metallic walls reflecting soft light, continuous yellow glowing floor line creating strong perspective and depth, vertical warm amber light panels on walls casting cinematic illumination, subtle sci-fi ambient fog in distance, high-tech yet elegant atmosphere. 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - From "lighting": Warm diffused lighting from vertical panels on both sides, soft golden-amber glow blending with cool metallic reflections, gentle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing three-dimensional form and skin texture, high dynamic range with vivid yet natural colors. 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - From "camera": Cinematic full-body shot with medium telephoto lens (85–135mm equivalent), f/2.0 aperture, razor-sharp focus from face to feet, shallow depth of field with creamy smooth bokeh on corridor background, slight low angle to emphasize power and grace, perfect viewpoint. 6. **Mood / atmosphere / emotion** - From "style" and "environment": elegant futuristic atmosphere with graceful and seductive vibe. 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - From "style": Ultra photorealistic cinematic photography. - From "quality": trending professional sci-fi fashion portrait. 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - From "quality": 8K resolution, extreme sharpness, rich cinematic colors with warm amber tones and cool metallic contrast, highly detailed skin realism, hair movement, outfit texture, corridor reflections and light play. - From "style" and general understanding: masterpiece quality. Now, assemble the prompt in the specified format. </think> A stunning young woman with European features and fair smooth skin showcasing hyper-detailed realistic textures, including visible fine pores, subtle peach fuzz, natural micro-imperfections, light freckles, sun-kissed glow on collarbones, athletic yet feminine body with realistic proportions, medium perky breasts with subtle natural cleavage visible through sheer sections of a form-fitting white and black tactical outfit, low-cut neckline accentuating toned midriff, slim waist, defined hips, long legs, long tousled blonde hair with sun-kissed highlights and individual strands gently moving, extremely detailed large brilliant blue-silver-golden-grey eyes with intricate iris textures and realistic reflections, full glossy lips transitioning into a warm radiant smile showing perfect white teeth, high cheekbones, soft natural blush on cheeks, dynamic confident walk down a futuristic corridor with hips swinging gracefully, one leg forward in mid-stride, arms relaxed yet purposeful, head tilted toward camera with engaging direct eye contact and growing smile, sleek futuristic corridor with metallic walls reflecting soft light, continuous yellow glowing floor line creating perspective and depth, vertical warm amber light panels on walls casting cinematic illumination, subtle sci-fi ambient fog in distance, high-tech yet elegant atmosphere, warm diffused lighting from vertical panels, soft golden-amber glow blending with cool metallic reflections, gentle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing form and skin texture, high dynamic range with vivid natural colors, cinematic full-body shot with medium telephoto lens (85–135mm equivalent), f/2.0 aperture, razor-sharp focus from face to feet, shallow depth of field with creamy smooth bokeh on corridor background, slight low angle emphasizing power and grace, perfect viewpoint, ultra photorealistic cinematic photography, maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture, realistic hair strands with natural movement, authentic fabric weave and subtle sheer effect, elegant futuristic atmosphere with graceful and seductive vibe, 8K resolution, extreme sharpness, rich cinematic colors with warm amber tones and cool metallic contrast, highly detailed skin realism, hair movement, outfit texture, corridor reflections and light play, trending professional sci-fi fashion portrait, masterpiece quality, ultra detailed, absurdres, intricate, sharp focus, depth of field, best quality, 8k uhd, hdr, photorealistic, realistic textures, dramatic rim lighting, volumetric fog, golden hour lighting, The subject is sharply depicted with outstanding detail., flux_pro
201525
Z Image
Ultra-realistic 8K scene of the person in the photo, stepping out of and breaking the panel border of a giant open manga book lying on the floor of an aesthetic bedroom. The comic pages depict a surreal 4-panel black-and-white manga with screentones, bold ink lines, and authentic Japanese layout, featuring the manga version of the same person trapped inside. The larger, real-life version emerges dynamically into the comic’s surface, vibrant, colorful, and photorealistic, in sharp contrast to the monochrome manga world. Both versions interact directly, creating a meta, humorous encounter. Highly detailed, crisp manga line art seamlessly integrated with vivid, realistic
432772
Nano Banana
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
6516
Nano Banana
{ "generation_request": { "meta_data": { "tool": "NanoBanana Pro", "task_type": "photobooth_contact_sheet_fashion_grid", "version": "v1.1_REFERENCE_IDENTITY_GRID_EN", "priority": "highest", "language": "en" }, "input": { "mode": "image_to_image", "reference_image_usage": "required", "notes": "Use the uploaded reference image to preserve the same person’s face, body features, hair, makeup and clothing. Maintain identity consistency in all frames." }, "output_settings": { "aspect_ratio": "4:5", "orientation": "portrait", "resolution_target": "ultra_high_res", "num_images": 1, "layout": { "type": "grid", "rows": 4, "cols": 4, "gutter": "thin_black_borders", "outer_border": "black_frame" }, "sharpness": "slightly_soft_flash", "film_grain": "medium_analog", "color_grade": "warm_2000s_flash_tone", "dynamic_range": "natural_not_hdr" }, "creative_prompt": { "scene_summary": "Early 2000s disposable camera photobooth contact sheet featuring the same person from the reference image in all 16 frames. Playful, candid, flash-lit aesthetic.", "environment": { "background": "plain white wall indoors", "lighting": "direct on-camera flash, slight hard shadow behind subject, subtle hotspot on skin, authentic point-and-shoot look" }, "subject": { "identity_consistency": "exact same person as reference image in all frames", "hair": "match reference image exactly", "makeup": "match reference image exactly", "wardrobe": "preserve the exact clothing from the reference image with accurate fabric texture and fit", "accessories": "keep accessories consistent with reference image" }, "poses_and_expressions": [ "hands on hips confident pose", "soft pout toward camera", "gentle smile", "hand on chin thoughtful look", "playful wink", "hands in hair", "slight shoulder turn", "looking off to the side", "serious model face", "soft laugh", "blowing a kiss", "tilted head cute expression", "crossed arms relaxed pose", "side glance over shoulder", "chin resting on hand", "natural relaxed expression" ], "style": { "aesthetic": "2000s photobooth disposable camera", "texture": "visible film grain, dust particles, tiny scratches, mild vignetting", "color": "slightly warm faded tones, subtle flash flattening", "lens": "35mm point-and-shoot softness" } }, "negative_prompt": [ "text", "logos", "watermark", "different outfit between frames", "different hairstyle", "different person", "face distortion", "identity drift", "overly polished editorial look", "HDR" ] } }
391
--- name: pastilles-llm-multi-agent description: Variante Claude Code du générateur de pastilles LLM, qui lance réellement plusieurs sous-agents en parallèle. Génère une pastille de communication interne sur les LLM (un texte court en français plus un prompt unique de génération d'images à coller dans le chat Gemini), via cinq sous-agents indépendants, une fusion, puis une revue critique par trois sous-agents et une correction. Utilise ce skill dès qu'on te demande de rédiger, produire ou générer une pastille, une fiche ou un contenu court d'acculturation sur les LLM, l'IA générative, le prompting, les agents, le RAG, la confidentialité IA ou tout sujet de la liste des 45 pastilles ci-dessous, que le mot "pastille" soit employé ou non. Utilise-le aussi dès qu'on te fournit un titre issu de cette liste. --- # Générateur de pastilles LLM, version multi-agents (Claude Code) ## Ce que fait ce skill Produit une pastille pédagogique complète à partir d'un titre. À la différence de la version chat, il lance de vrais sous-agents parallèles: une passe de recherche, puis cinq sous-agents indépendants qui rédigent chacun un brouillon sous un angle différent, puis une fusion par l'orchestrateur qui retient les meilleures formulations, et enfin une revue critique par trois sous-agents suivie d'une correction. Livrable: le texte de la pastille et un prompt unique de génération d'images à coller dans le chat Gemini. ## Environnement requis Ce skill suppose Claude Code, avec les sous-agents (outil Task) et la recherche web disponibles. Si les sous-agents ne sont pas disponibles dans ton environnement, n'improvise pas: utilise la variante chat, qui fait le même travail en self-ensemble séquentiel. ## Entrée Un titre de pastille, idéalement issu de la liste ci-dessous. Si le titre est ambigu ou hors liste, demande une clarification avant de générer. ## Liste des 45 pastilles (pour la continuité) 1. Au fait, c'est quoi un LLM ? 2. Modèles de fondation : ces géants pré-entraînés dont tout découle 3. Dans les coulisses : comment une IA apprend (sans vraiment comprendre) 4. Le contexte : la "mémoire vive" de l'IA (ce qu'elle voit à un instant T) 5. Pourquoi l'IA oublie : la vérité sur la mémoire d'une session à l'autre 6. Température et créativité : régler le curseur de la "folie" de l'IA 7. Les tokens : la monnaie d'échange (et la manière de penser) des LLM 8. Le non-déterminisme : pourquoi l'IA fait des stats et ne donne jamais deux fois la même réponse 9. Multimodalité : quand l'IA commence à voir, entendre et coder 10. Knowledge cutoff : pourquoi votre IA vit (parfois) dans le passé 11. Les biais de l'IA : le miroir pas toujours glorieux de nos propres données 12. Zéro pointé : pourquoi l'IA est structurellement nulle en calcul mental 13. Les hallucinations : quand l'IA invente la vie avec un aplomb incroyable 14. L'art du Fact-Checking : comment ne pas gober les hallucinations de l'IA 15. L'empreinte carbone de l'IA : combien d'eau et de CO2 pour un prompt ? 16. Anatomie d'un bon prompt : la recette de base 17. Les pièges du prompt : flou artistique et overdose d'instructions 18. L'art de la discussion : pourquoi il faut itérer plutôt que tout demander d'un coup 19. Donner des ordres : comment imposer un format de sortie strict (JSON, tableaux, etc.) 20. Diviser pour régner : découper une tâche complexe en étapes claires 21. Le Few-Shot Prompting : un bon exemple vaut mille consignes 22. Inception : donner un rôle et un contexte à votre IA 23. Zéro syndrome de la page blanche : rédiger et reformuler ses mails ou CR 24. La diagonale du fou : résumer et synthétiser un pavé de 50 pages en 2 minutes 25. Le sparring-partner : brainstormer et structurer ses idées sans filtre 26. Caméléon : traduire et adapter son ton selon le destinataire 27. Data-analyst de poche : faire parler un tableur ou des données brutes 28. Mise en situation : préparer une réunion ou un brief client avec l'IA 29. Avis aux dévs : l'IA comme copilote de code et de review 30. SLM : Pourquoi sortir un tank pour écraser une mouche ? 31. Le "harnais" : donner des bras, des yeux et un cadre au LLM 32. C'est quoi un "Agent" IA, concrètement ? 33. Human-in-the-loop : pourquoi il faut toujours garder un humain aux manettes 34. Quand l'agent s'emballe : autonomie et erreurs en cascade 35. Automatisation : confier un workflow complet à un agent 36. Tool Use : quand l'IA apprend à utiliser une calculatrice ou une API 37. Le RAG : faire réviser vos propres documents à l'IA avant qu'elle ne réponde 38. MCP et connecteurs : brancher l'IA directement sur vos outils de dev et d'orga 39. Responsabilité : vous êtes le seul signataire de ce que produit l'IA 40. Confidentialité : où vont vraiment les données que vous tapez ? 41. Alerte Rouge : les données clients à ne JAMAIS copier-coller dans une IA public 42. Shadow AI : pourquoi le ChatGPT gratuit perso au bureau est une fausse bonne idée 43. RGPD et IA : ce qu'il faut savoir pour rester dans les clous 44. Prompt Injection : quand on pirate une IA avec de simples phrases 45. La Chaîne de Pensée : pourquoi l'IA a besoin de "réfléchir à voix haute" Sers-toi de cette liste pour situer la pastille et délimiter son périmètre. Avant de rédiger: - Repère les 1 à 3 pastilles voisines qui recouvrent le sujet, et dresse une courte liste "déjà traité ailleurs, à ne pas ré-expliquer". - Si une pastille voisine a déjà été produite, demande son texte à l'utilisateur (ou appuie-toi dessus s'il est fourni) pour caler précisément la limite. - Les concepts fondateurs déjà couverts ne sont rappelés qu'en une phrase de mise en contexte, jamais ré-expliqués. Le coeur de la pastille est consacré à ce que son titre ajoute de neuf. - Ne renvoie pas vers les autres pastilles dans le texte final, sauf si le titre l'impose. ## Règles du texte - Longueur adaptée à la profondeur du sujet. Un sujet léger tient en 3 paragraphes courts; un sujet plus riche peut aller jusqu'à 4 paragraphes plus étoffés. Ne gonfle pas artificiellement un sujet simple et ne compresse pas à l'excès un sujet dense: juge la profondeur par la richesse réelle du concept. Prose uniquement, pas de listes ni de puces. - Ton décontracté, précis et léger: accessible, vivant et sans lourdeur, mais techniquement juste. Public visé: une main-d'oeuvre diverse en société de conseil, du profil non technique au développeur. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Reste léger, évite le name-dropping. N'accumule pas les noms de modèles, d'outils, d'entreprises, de chercheurs ou de techniques pour faire savant: privilégie l'explication claire du mécanisme. Ne cite un nom précis que s'il éclaire vraiment le propos, une ou deux fois au maximum. La recherche sert l'exactitude et alimente la section Sources, elle n'a pas à truffer le texte de références. - Contenu autonome: la pastille se comprend seule, sans avoir lu les autres, sauf mention contraire dans le titre. - Rédige en français. - Ne mentionne jamais ANEO ni aucun nom d'entreprise. - Le texte explicatif ne figure jamais dans l'image. - N'utilise pas de tiret cadratin. Privilégie des caractères standard (virgules, deux-points, parenthèses). ## Processus multi-agents ### Étape 1, recherche (orchestrateur, une seule fois) Toi, l'orchestrateur, fais d'abord une passe de recherche web ciblée sur le sujet de la pastille. Une seule passe, en amont. Les sous-agents ne peuvent pas lancer de sous-agents et n'héritent pas de ton contexte, donc c'est à toi de faire la recherche puis de leur transmettre les résultats. - Objectif: exactitude, chiffres, exemples concrets et faits à jour. - Ancre tes recherches dans le présent: repère la date du jour fournie dans ton contexte (champ currentDate) et privilégie les informations et chiffres les plus récents, car ce domaine évolue vite. Précise cette date au besoin dans tes requêtes. - Priorise les sources originales ou officielles (blogs d'éditeurs, articles de recherche, sites gouvernementaux, documentation) plutôt que les agrégateurs. - Adapte la profondeur: quelques recherches pour un concept stable (par exemple ce qu'est un token), davantage pour un sujet à chiffres ou mouvant (empreinte carbone et coûts, modèles spécifiques, SLM, MCP et connecteurs, RGPD). - Produis un brief de recherche compact: faits clés, chiffres utiles, et 2 à 4 sources. Tu l'injecteras tel quel dans chaque sous-agent, et tu listeras les sources dans le livrable final. ### Étape 2, affinement de la thématique (orchestrateur, une seule fois) Si plusieurs axes apparaissent pertinents après avoir enlever les axes adressés par d'autres pastilles prévues, dialogue avec l'utilisateur pour choisir l'axe de la pastille. Si l'axe choisi nécessite de nouvelles sources, refait en une recherche web selon les consignes précédentes. ### Étape 3, fan-out (cinq sous-agents en parallèle) Lance cinq sous-agents via l'outil Task, dans le même tour, pour qu'ils s'exécutent en parallèle. Sois explicite sur le parallélisme: par défaut Claude Code reste séquentiel, il ne parallélise que si tu le demandes clairement. Un sous-agent par angle: 1. Analogie: expliquer le concept via une métaphore concrète du quotidien. 2. Cas d'usage: partir d'une situation de travail réelle et vécue. 3. Idée reçue: démarrer d'un malentendu courant, puis rectifier. 4. Mécanique: aller au coeur du fonctionnement, précis mais accessible. 5. Enjeu: pourquoi ça compte, ce que ça change concrètement pour l'utilisateur. Comme le contexte n'est pas hérité, chaque prompt de sous-agent doit être autonome et contenir tout le nécessaire. Utilise ce gabarit, en remplaçant les crochets: ``` Tu rédiges un brouillon d'une pastille pédagogique interne sur les LLM. Tu travailles seul, sans accès au reste de la conversation. Titre de la pastille: [TITRE] Angle imposé: [ANGLE, par exemple "Analogie: expliquer via une métaphore concrète du quotidien"] À NE PAS ré-expliquer (déjà couvert par d'autres pastilles), à mentionner en une phrase au maximum: [LISTE des points déjà traités ailleurs, à ne pas retraiter] Brief de recherche (appuie-toi dessus, ne relance pas de recherche): [BRIEF: faits clés, chiffres, sources] Règles: - Longueur selon la profondeur du sujet: 3 paragraphes courts si le sujet est léger, jusqu'à 4 paragraphes étoffés s'il est dense. Prose uniquement, pas de listes. - Mets en gras et/ou en italique les termes importants, de l'ordre de 3 à 4 emphases par paragraphe (repère indicatif, ne force pas ce compte). - Ton décontracté, précis et léger, accessible mais techniquement juste. Évite le name-dropping: pas d'accumulation de noms de modèles, outils, entreprises ou chercheurs, privilégie l'explication du mécanisme. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Français. Ne mentionne aucune entreprise. Pas de tiret cadratin. - Contenu autonome, compréhensible seul. Réponds exactement dans ce format: TEXTE: [les 3 à 4 paragraphes] ILLUSTRATION_TITRE: [une phrase décrivant le concept central à illustrer] SCHEMA: [oui ou non. Si oui, décris en 2 à 3 lignes le diagramme et ses libellés en français] ``` ### Étape 4, fan-in et fusion (orchestrateur) Attends les cinq retours, puis fusionne en une seule pastille finale: - Retiens la structure la plus claire, la meilleure analogie, l'exemple le plus parlant, la correction la plus nette, la formulation la plus précise et le "pourquoi ça compte" le plus fort. - Réécris en une seule voix cohérente. Pas d'effet patchwork. - Respecte la longueur adaptée à la profondeur (voir Règles du texte) et le ton décontracté, précis et léger, sans name-dropping. - Décide s'il faut un schéma: inclus-en un si au moins deux sous-agents ont répondu "oui", ou si le concept est intrinsèquement un processus, un flux ou une comparaison (par exemple RAG, agents, étapes de prompting, comparaison de modèles). Le cas échéant, fusionne les meilleures idées de schéma. - Construis ensuite le prompt image unique (section ci-dessous). ### Règles d'écriture pour la pastille finale - Longueur selon la profondeur du sujet: 3 paragraphes courts si le sujet est léger, jusqu'à 4 paragraphes étoffés s'il est dense. Prose uniquement, pas de listes. - Mets en gras et/ou en italique les termes importants, de l'ordre de 3 à 4 emphases par paragraphe (repère indicatif, ne force pas ce compte). - Ton décontracté, précis et léger, accessible mais techniquement juste. Évite le name-dropping: pas d'accumulation de noms de modèles, outils, entreprises ou chercheurs, privilégie l'explication du mécanisme. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Français. Ne mentionne aucune entreprise. Pas de tiret cadratin. - Contenu autonome, compréhensible seul. ### Étape 5, revue critique et correction (trois sous-agents en parallèle, puis orchestrateur) Quand la lancer: systématiquement à la première génération, une fois le texte fusionné et le prompt image construits. Ensuite, si l'utilisateur demande des ajustements, ne relance pas la revue d'office: propose-la, et ne la lance qu'avec son accord (elle coûte trois sous-agents de plus). Principe: les relecteurs produisent des CONSTATS, jamais une réécriture. Chacun renvoie une liste de problèmes localisés, assortis d'une gravité et d'un correctif précis. C'est toi, l'orchestrateur, qui appliques et réécris, pour garder une voix unique et éviter l'effet patchwork. Ce que la revue peut juger: les relecteurs ne voient pas les images, générées plus tard dans Gemini. La revue du visuel porte donc uniquement sur le PROMPT image (clarté, cohérence avec le texte, respect des consignes), jamais sur un rendu. Le contrôle visuel réel reste à l'utilisateur. Lance trois sous-agents en parallèle (parallélisme explicite, sinon l'exécution sera séquentielle), un par grille. Le contexte n'étant pas hérité, chaque prompt doit être autonome et contenir tout le nécessaire: le titre, le texte fusionné, le bloc prompt image, le brief de recherche, la liste "déjà traité ailleurs" (et les textes des pastilles voisines s'ils sont disponibles), et la liste des 45 titres pour repérer les chevauchements. Les trois grilles: 1. Fond, exactitude et périmètre: exactitude vs le brief (aucun chiffre, date ou fait inventé ni sur-affirmé, rien qui le contredise), cohérence entre le titre et ce que le texte délivre, chevauchements avec les pastilles voisines (le texte ré-explique-t-il ce qui est traité ailleurs ? redites à signaler), autonomie du contenu, clarté du message à retenir, repérage de ce qui vieillira mal (à nuancer). 2. Forme, ton et pédagogie: ton décontracté-précis-léger et techniquement juste, absence de name-dropping, rythme et fluidité à la lecture à voix haute, longueur adaptée à la profondeur (3 à 4 paragraphes, prose, pas de listes) et chasse au verbiage, accessibilité pour un profil non technique (jargon expliqué, analogies claires), présence d'emphases utiles (gras/italique), de l'ordre de 3 à 4 par paragraphe à titre indicatif (signale un texte trop peu emphasé autant qu'un excès, mais ne fais jamais retirer une emphase pertinente), force de l'accroche. 3. Conformité et visuel: contraintes dures (aucun tiret cadratin ni caractère non standard, aucun nom d'entreprise ni ANEO, français correct, prose sans listes ni puces); puis le prompt image: titre exact au caractère près, charte présente une seule fois, texte de la pastille bien marqué "à ne pas afficher", illustration-titre iconique et non schéma de processus, schéma en 2e image séparée si retenu, libellés de schéma courts (groupes nominaux) et en français, cohérence entre le visuel décrit et le coeur du texte. Gabarit de relecteur (remplace les crochets): ``` Tu relis un brouillon quasi final de pastille pédagogique interne sur les LLM. Tu travailles seul, sans accès au reste de la conversation. Tu ne réécris pas: tu rends des constats et des correctifs précis. Titre de la pastille: [TITRE] Texte fusionné à relire: [TEXTE COMPLET] Prompt image associé (à relire aussi): [BLOC PROMPT IMAGE] Brief de recherche de référence (fait foi pour l'exactitude): [BRIEF] Déjà traité par d'autres pastilles, à ne pas ré-expliquer (et textes voisins si fournis): [LISTE + TEXTES VOISINS] Liste des 45 titres (pour repérer les chevauchements): [LISTE DES 45 TITRES] Ta grille de relecture (tiens-t'en strictement à elle): [RUBRIQUE 1, 2 OU 3] Rappel: tu ne peux pas voir d'image; juge seulement le PROMPT image, pas un rendu. Réponds exactement dans ce format: CONSTATS: - [bloquant|recommandé|mineur] [problème localisé] -> [correctif précis] - ... (n'invente pas de problème: si un point est bon, ne le liste pas) VERDICT: [une phrase: publiable tel quel / corrections mineures / corrections nécessaires] ``` Fan-in et correction (orchestrateur): rassemble les constats, dédoublonne, arbitre les conseils contradictoires. Garde-fou anti-gonflement: entre "ajouter" et "raccourcir", la concision l'emporte, sauf erreur de fond avérée. Applique les constats bloquants et recommandés, écarte ou mentionne les mineurs. Une seule passe, pas de boucle. Réécris la version corrigée en une seule voix, puis prépare le court résumé "Ce que la revue a corrigé" (voir Format de sortie). ### Points de vigilance Claude Code - Parallélisme explicite: demande clairement cinq sous-agents en parallèle, sinon l'exécution sera séquentielle. - Contexte non hérité: mets tout dans le prompt du sous-agent, en particulier le brief de recherche. - Pas d'imbrication: un sous-agent ne peut pas en lancer un autre, garde recherche et fusion chez l'orchestrateur. - Coût et limites: cinq sous-agents en parallèle multiplient l'usage de jetons et peuvent déclencher des limites de débit. En cas de limite, replie-toi sur une exécution séquentielle. - Revue: la phase de revue ajoute trois sous-agents. Elle tourne d'office à la première génération; sur les demandes d'ajustement ultérieures, propose-la et ne la lance qu'avec l'accord de l'utilisateur. En cas de limite de débit, exécute les relecteurs en séquentiel ou replie-toi sur un relecteur global unique. - Réemploi optionnel: tu peux définir les cinq angles comme sous-agents personnalisés dans .claude/agents/ pour les réutiliser, mais ce n'est pas obligatoire, les sous-agents polyvalents suffisent. ## Prompt de génération d'images (un seul bloc, à coller dans le chat Gemini) Produis un seul bloc de prompt, rédigé en français, à coller tel quel dans le chat de Gemini. Gemini se charge d'appeler Nano Banana une ou deux fois, ce qui réduit les manipulations pour l'utilisateur. Le bloc demande: - Toujours une illustration-titre. - Un schéma seulement s'il a été retenu à l'étape de fusion, généré comme une seconde image séparée (fichier distinct), jamais dans la même image que le titre. Contenu du bloc: - Décris la charte graphique une seule fois (bloc ci-dessous) et précise que toutes les images la partagent, pour une cohérence visuelle. - Précise: tout texte affiché dans les images est en français. Cette consigne est importante, elle évite qu'un libellé se retrouve en anglais. - Inclus le texte complet de la pastille dans le bloc, en contexte de génération, clairement marqué comme à ne pas afficher. Il aide le modèle à comprendre le sujet et à choisir une illustration juste. Les seuls textes rendus visibles sont le titre exact et, pour un schéma, ses libellés. Le texte de la pastille ne doit jamais apparaître dans les images. - Illustration-titre: composition épurée et moderne, focus graphique central iconique représentant le concept. Ce n'est pas un schéma de processus. Le seul texte affiché est le titre exact, en en-tête: n'ajoute aucun sous-titre, accroche ou texte secondaire, le titre seul. Technique en deux temps pour le titre: dans le prompt généré, écris le titre exact entre guillemets droits (forme attendue: Le titre exact: "...") et demande un rendu fidèle, sans faute, en police sans serif corporate. Les guillemets ne sont qu'un délimiteur côté prompt: ne les commente pas et n'ajoute aucune consigne à leur sujet dans le prompt généré (pas de mention du type "les guillemets ne doivent pas apparaître"). - Schéma s'il est retenu: une seconde image générée séparément, dans son propre fichier, jamais fusionnée avec l'illustration-titre ni juxtaposée dans la même image. Diagramme d'entreprise propre et net (processus, flowchart ou comparaison), cohérent avec l'illustration-titre, mêmes charte et style, libellés fournis explicitement en français, très peu de fioritures. - Libellés de schéma: des groupes nominaux courts et lisibles seuls (par exemple "Exemples de bonnes réponses"), pas des phrases verbales qui sonnent comme des ordres ("Montrer de bonnes réponses"). Quelques mots par libellé. Attention aux titres longs: Nano Banana rend fiablement les libellés courts mais peut faire une faute sur un titre long; la technique en deux temps limite le risque. Si le rendu d'un titre long reste incertain, indique à l'utilisateur qu'il peut demander à Gemini d'utiliser Nano Banana Pro, plus fiable pour le texte long. Gabarit, illustration-titre seule (exemple avec le titre 1): ``` Génère une image en respectant cette charte graphique: [bloc Charte graphique]. Contexte pour comprendre le sujet, à NE PAS afficher dans l'image: [ici le texte complet de la pastille]. Illustration-titre pour une pastille pédagogique sur les LLM. Composition épurée et moderne, focus graphique central iconique: un nuage de lettres et de mots qui se condense en une bulle de dialogue, évoquant un modèle de langage qui transforme du texte en réponse. Ce n'est pas un schéma de processus. Seul texte à afficher dans l'image: le titre, en en-tête, sans faute d'orthographe, en police sans serif corporate, sans sous-titre ni texte secondaire. Le titre exact: "TITRE EXACT". Tout le texte affiché est en français. Format 16:9. ``` Gabarit, illustration-titre plus schéma (deux images): ``` Prépares-toi à générer deux images séparées, dans deux fichiers distincts. Les deux partagent cette charte graphique: [bloc Charte graphique]. Tout texte affiché dans les images est en français. Contexte pour comprendre le sujet, à NE PAS afficher dans les images: [ici le texte complet de la pastille]. Image 1, illustration-titre: composition épurée et moderne, focus graphique central iconique qui illustre le sujet. Ce n'est pas un schéma de processus. Seul texte à afficher: le titre, en en-tête, sans faute d'orthographe, police sans serif corporate, sans sous-titre ni texte secondaire. Le titre exact: "TITRE EXACT". Format 16:9. Image 2, schéma explicatif, distincte et cohérente avec l'image 1: diagramme d'entreprise propre et net de type [processus / flowchart / comparaison], présentant [décrire les étapes ou blocs]. Libellés exacts à afficher en français: [liste des libellés]. Très peu de fioritures, focus sur la clarté, n'inclus pas le titre. Génère dans un premier temps seulement la première image (l'illustration-titre), et attends les instructions de l'utilisateur pour générer la 2e image. ``` ## Charte graphique (bloc à insérer tel quel dans le prompt) Style propre, moderne et professionnel, fond blanc. Palette: orange vif et bleu corporate foncé en dominantes, blanc et gris pour les respirations, accents multiculturels discrets (rouge, vert, jaune, bleu) reprenant subtilement un motif de couleurs de logo. Superpositions graphiques épurées: motifs géométriques abstraits, quartiers de cercle, lignes claires. Composition aérée, jamais surchargée. Typographie sans serif, propre et corporate. ## Format de sortie Le travail des sous-agents reste dans leur propre contexte, seuls leurs retours te reviennent. N'affiche donc que le livrable final (déjà corrigé par l'étape de revue), la conversation principale reste propre. Si l'utilisateur demande à voir l'atelier, restitue les cinq brouillons reçus et les constats de revue. Livrable final, dans cet ordre: - Quand une revue a eu lieu (systématiquement à la première génération), un court résumé "Ce que la revue a corrigé" (2 à 4 lignes) présentant les principaux ajustements, avant le reste. Lors d'ajustements ultérieurs sans revue, présente simplement la version ajustée sans ce résumé. - Le texte de la pastille (3 à 4 paragraphes), dans sa version corrigée. - Un bloc de code intitulé "Prompt images (à coller dans Gemini)", contenant le prompt unique prêt à copier. - Une courte section "Sources" listant 2 à 4 références principales issues de l'étape de recherche, de préférence officielles ou originales. Cette section sert à la vérification et n'a pas vocation à être publiée dans la pastille. Termine toujours par cette question exacte: "Comment trouvez-vous le texte, l'image titre et le diagramme (si généré) ? Si une partie vous semble trop complexe, ou si vous souhaitez affiner le focus d'un visuel pour qu'il soit encore plus épuré, n'hésitez pas à me le faire savoir, et je l'ajusterai." -> le personnage doit incarner le professeur un peu foufou qui allusine et obsédé par l'ai
015
Grok Image
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
507522
Nano Banana
Learn to create better images with AI! Check out our crash course in prompt engineering & AI art generation! AM1970 posted about 2 hours ago 8 views 0 comments Behold this realistic masterpiece: a slim, elegant woman with fiery red hair tied in a long ponytail, defined eyebrows framing her striking blue eyes, and pale skin dotted with charming freckles across her face and chest. Captured from a side angle, she radiates happiness with a closed-mouth smile, her upturned nose adding to her allure. Dressed in a vibrant, colorful fabric robe pieced together with zips, buttons, seams, and intricate stitching—fully clothed yet teasing partial nudity with subtle cleavage and teardrop-shaped, saggy breasts for a natural, confident vibe. Illuminated by dramatic rim lighting that highlights every textured, detailed inch of her perfect skin, this high-quality image is a celebration of beauty, realism, and artistry! Read more... Comments (0) Add a comment Stunning handsome readhead woman by AM1970 • Follow 0 realistic, masterpiece, high quality BREAK 1woman, 28 y. o. slim fit woman, red hair, long pony tail, defined eyebrows, (fully clothed) , pale skin, (realistic skin:1.3) | (textured skin:1.3) | (detailed skin:1.3) | (perfect skin:1.3) , (upturned nose) , freckles on face and chest, colorful fabric combining to form robe zips, buttons, seams, stitching), partial nudity, teardrop boobs, blue eyes, (cleavage:1.2) , saggy breasts. side angle. happy| closed mouth. ((rim lighting) ) Negative prompt: unrealistic, saturated, high contrast, big nose, painting, drawing, sketch, cartoon, anime, manga, render, CG, 3d, watermark, signature, label, nsfw, nudity, fetish, latex, pvc, suggestive pose, weapon, combat, blood, war scene, aggressive expression, harsh shadows, heavy makeup, color tint, saturation, chromatic aberration, lens flare, low resolution, blur, noise, artifacts, over-retouching, plastic skin, oversharpening, watermark, text, logo Generation parameters 832x1216 Model used CyberRealistic XL v7.0 [Download AI model] More by AM1970 Twitter Instagram Say hi 👋 PromptHero is built by ElevenYellow . Navigation Create AI images Academy - Learn AI PromptHero Community Affiliate program Openjourney AI Challenges Verified profiles Plans Advertise FAQ 🟢 Status Learn AI Content Creation with AI for Brands and Products Midjourney for graphic design & art professionals Crash course in generative AI & prompt engineering for images AI influencers and consistent characters Create custom AI models and LoRas by fine-tuning Stable Diffusion Master your composition: advanced AI image generation with ControlNet Guides Best AI Upscalers 2025 Best AI Video Editors 2025 Best AI 3D Models 2025 Best AI Image Editors 2025 Best AI Notetakers 2025 The Best AI Prompts Best AI Prompts Best Veo Prompts Best Hailuo Prompts Best Wanx Prompts Best Hunyuan Prompts Best Seedance Prompts Best Pixverse Prompts Best Luma Prompts Best Kling Prompts Best Runway Prompts Best Nano Banana Prompts Best SeedEdit Prompts Best Qwen Prompts Best FLUX Prompts Best Recraft Prompts Best Ideogram Prompts Best Seedream 4 Prompts Best Google Imagen Prompts Best ChatGPT Image Prompts Best Stable Diffusion Prompts Best DreamShaper Prompts Best Realistic Vision Prompts Best epiCRealism Prompts Best majicMIX realistic Prompts Best Midjourney Prompts Best Openjourney Prompts Best ChilloutMix Prompts Best Deliberate Prompts Best ReV Animated Prompts Best Perfect World Prompts Best Dreamlike Photoreal Prompts Best Dreamlike Diffusion Prompts Best Anything Prompts Best Protogen Prompts Best AbyssOrangeMix Prompts Best Grapefruit Prompts Best Kenshi Prompts Best Analog Diffusion Prompts Best SynthwavePunk Prompts Best Illuminati Diffusion Prompts Best Arcane Diffusion Prompts Best Mo-Di Diffusion Prompts Best Redshift Diffusion Prompts Best Waifu Diffusion Prompts Best TrinArt Prompts Best Poolsuite Diffusion Prompts Best Funko Diffusion Prompts Best Text to Pokémon Prompts Best DALL-E Prompts Best Fashion Diffusion Prompts AI Models All AI models Image generation AI Models Video generation AI Models Large Language Models LoRA Textual Inversion ControlNet Hypernetwork Aesthetic Gradient LyCORIS VAE Poses Wildcards Workflows Other Other sites Dreamshot Jobician 2025 © PromptHero @prompthero
1226
FLUX
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
5393
Nano Banana
Ultra-realistic 8K scene of the person in the photo, stepping out of and breaking the panel border of a giant open manga book lying on the floor of an aesthetic bedroom. The comic pages depict a surreal 4-panel black-and-white manga with screentones, bold ink lines, and authentic Japanese layout, featuring the manga version of the same person trapped inside. The larger, real-life version emerges dynamically into the comic’s surface, vibrant, colorful, and photorealistic, in sharp contrast to the monochrome manga world. Both versions interact directly, creating a meta, humorous encounter. Highly detailed, crisp manga line art seamlessly integrated with vivid, realistic
24964
Nano Banana
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
7821
Nano Banana
<think> The user wants a Z-Image-Turbo prompt based on the provided description. I need to create a prompt following the specified structure and format, extracting information from the user's request and transforming it into a high-quality prompt for image generation. Let's break down each section of the prompt based on the user's input: 1. **Subject + very detailed description (who/what, appearance, clothing, expression, pose, age, ethnicity if relevant)** - Subject: beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs - Appearance: porcelain skin, fair smooth skin showing hyper-detailed realistic texture: visible fine pores, subtle peach fuzz on shoulders and upper back, natural micro-imperfections, light natural freckles across shoulders and upper spine, lightly tanned sun-kissed glow on back and arms - Clothing: elegant emerald silk dress (but the description focuses more on sheer white translucent curtains and lace details) -> delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong - Expression: intimate, slightly parted-lips expression, serene and contemplative expression - Pose: three-quarter back view, head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding the curtains apart - Ethnicity: Caucasian woman 2. **Action or scene context** - Intimate, sensual mood, subtle NSFW touch 3. **Environment / setting + time of day + weather** - Bright boudoir bedroom with large window behind curtains, faint white flowers in soft-focus background, polished wooden floor partially visible, luxurious yet minimal setting - Time of day: bright soft diffused natural daylight (implies daytime) 4. **Lighting (extremely important for this model – always describe dramatic, cinematic, volumetric, or natural lighting in detail)** - Bright soft diffused natural daylight from large window - High-key ethereal mood with gentle gradients and soft wrap-around light - Subtle rim light outlining hair, shoulders and curves - Realistic soft shadows enhancing skin texture and lace details 5. **Camera angle, lens, shot type (e.g. close-up, wide shot, low angle, 35mm, 85mm, cinematic)** - Vertical portrait composition, shot on high-end full-frame camera with 85mm lens - f/1.8 aperture, razor-sharp focus on face, eyes, hair, skin pores and lace texture 6. **Mood / atmosphere / emotion** - Ethereal dreamy atmosphere - Intimate sensual mood - Serene and contemplative expression 7. **Art style influences (photorealistic, hyperrealistic, cinematic, film still, National Geographic style, etc.)** - Ultra photorealistic cinematic boudoir photography - Maximum lifelike skin detail with pores, fine body hair, natural imperfections and sunlit texture 8. **Technical quality boosters (masterpiece, best quality, ultra-detailed, 8k, sharp focus, intricate details, highly detailed skin, realistic textures)** - 8K resolution - extreme sharpness - rich natural colors with soft pastel lavender and warm skin tones - high dynamic range - highly detailed skin realism - hair movement - curtain folds - lace patterns and light diffusion - trending professional boudoir portrait - masterpiece quality Now, let's construct the prompt following the exact structure and incorporating all these details into a coherent text. </think> A breathtaking ultra-realistic vertical portrait of a beautiful young Caucasian woman with short dark red messy bob hair and choppy bangs, delicate pale lavender lace bra with intricate floral pattern, matching ultra-thin string lace thong revealing toned lower back, hips and round firm buttocks for a subtle NSFW touch, standing three-quarter back view with head turned over shoulder, soft arch in lower back emphasizing natural spine curve and hip shape, relaxed yet seductive posture, weight shifted naturally to one leg, arms raised gently holding sheer white translucent curtains apart, sheer fabric clinging slightly to skin creating soft folds and transparency, scene bathed in bright soft diffused natural light from a large window behind curtains, polished wooden floor partially visible, faint white flowers in soft-focus background, luxurious yet minimal setting, high-key ethereal mood with gentle gradients and soft wrap-around light, subtle rim light outlining hair, shoulders and curves, realistic soft shadows enhancing skin texture and lace details, vertical portrait composition shot on high-end full-frame camera with 85mm lens f/1.8 aperture razor-sharp focus on face eyes hair skin pores and lace texture shallow depth of field with creamy dreamy bokeh on curtains and background flowers ultra photorealistic cinematic boudoir photography maximum lifelike skin detail with pores fine body hair natural imperfections and sunlit texture realistic lace weave and transparency ethereal dreamlike atmosphere intimate sensual mood without explicit exposure masterpiece quality 8K resolution extreme sharpness rich natural colors with soft pastel lavender and warm skin tones high dynamic range highly detailed skin realism hair movement curtain folds lace patterns and light diffusion trending professional boudoir portrait., The subject is sharply depicted with outstanding detail., flux_pro
3246
Nano Banana
High-definition quality, photo version. Shot with a Canon EOS 5D Mark IV and an EF 100mm f/2.8L Macro IS USM lens, f/4.5, 1/200s, ISO 400. Best quality, ultra HD, 8K, award-winning photography. --ar 3:4 High-definition, realistic skin textures, extremely detailed facial textures, cinematic lighting, and a natural atmosphere, no cartoon style.
10353
Sora
Steps: 20, Sampler: DPM++ 2M Karras, CFG scale: 7, Seed: 413127992, Size: 1280x1600, Model hash: 7fa552ffb0, Model: nijidiffusedmix_v40, Denoising strength: 0.3, Clip skip: 2, ADetailer model: face_yolov8s.pt, ADetailer prompt: "dark blue eyes, freckles below eyes", ADetailer confidence: 0.3, ADetailer dilate erode: 4, ADetailer mask blur: 4, ADetailer denoising strength: 0.5, ADetailer inpaint only masked: True, ADetailer inpaint padding: 32, ADetailer version: 23.12.0, Version: v1.6.1
10493
Stable Diffusion
((abstract illusionism)), 8k full wide focus, (Masterpiece:1.2), ((surreal)),cinematic still, ((photorealistism)), (psychedelic imagery), trending on artstation, sharp focus, studio photo, intricate details, highly detailed, by greg rutkowski In the ethereal realm of the surreal dreamscape, envision a mesmerizing painting that beckons viewers into a kaleidoscope of fantastical wonders, a realm where imagination knows no bounds Negative prompt: nsfw, nude,blurry, low quality, oddly shaped, disfigured,unclear || undefined || mutations || deformities || off center || poor_composition Steps: 100, Sampler: DPM++ 3M SDE, CFG scale: 14.0, Seed: 33, Face restoration: CodeFormer, Size: 1280x832, Model hash: 1895a00bfc, Model: SSD-1B-A1111, Denoising strength: 0.31000000000000005, Mask blur: 4, Noise multiplier: 1.05, Version: v1.6.0-2-g4afaaf8a
0132
Stable Diffusion
Among the growing list of AI photo editing tools, Deepfake Maker takes a slightly different approach. While it’s best known for facial swaps and animated transformations, it also includes a surprisingly capable AI clothing removal feature. It’s not for everyone, but it’s an interesting choice for users looking for more dramatic or playful results. If you're after a quick way to generate bold clothing edits or outfit swaps, Deepfake Maker might surprise you. It’s not the most refined tool out there, but it’s definitely one of the most entertaining. 4. PxBee PxBee Source: PxBee Available On: Web Browser Price: Free *Annual subscription starts at $00.27/image and you can choose to download 50, 125, or 750 images monthly per year Pros Decent auto-detection for basic clothing removal Integrated with other tools Cons Not suitable for high-end projects Results take longer than expected to generate PxBee is better known for its stock photo collection and image generation features, but it also offers a basic AI clothing removal function built into its photo editing toolkit. While it's not as feature-rich or advanced as some dedicated AI editors, its simplicity and accessibility make it a decent option for casual users who want fast, no-frills results. PxBee may not be the most powerful AI clothing remover out there, but it's a solid pick for quick edits and simple concept visuals. Its intuitive UI and browser-based access make it appealing for users who want fast results without the learning curve. If you're after something basic, affordable, and easy to use, it's a tool worth exploring. 5. AI Face Swap AI Face Swap Source: AI Face Swap Available On: Web Browser Price: $1 per image Pros No manual masking needed Prompt-based customization Presets for inspiration Cons No fine-tuning or editing control With AI Face Swap, you can enter prompts like "red bikini", "formal suit", or even "nude", and the AI will generate a realistic version of your uploaded image that reflects your description. You can also choose from quick preset styles like bikini, romper, or skirt for faster results. This level of control makes the tool especially appealing to users looking for stylized edits, fantasy concepts, or mockup visuals. It also eliminates the need for manual masking or precise editing. 6. Bylo AI BYLO AI Source: Bylo.ai Available On: Web Browser Price: Free *Premium subscription available from $9.99 Pros Prompt-based clothing edits Free trial credits Part of a growing creative suite Cons No manual adjustment tools Limited free usage Relies heavily on prompts Bylo AI is a user-friendly AI platform that offers a range of creative features, including an AI Clothes Remover. The tool is built with simplicity in mind: just upload a photo, enter your prompt, and let the AI generate a clean, edited version. Bylo AI’s Clothes Remover is a great choice if you're looking for a lightweight, affordable tool that lets you experiment with clothing changes using just a text prompt. It won’t replace professional retouching tools, but for concept edits, social media content, or fun transformations, it delivers exactly what most casual users need. FAQs About AI Clothing Remover & Replacer 1. How do I remove clothes in photos? 2. What is the best online tool to remove clothes? 3. Are AI clothes removers safe?
31317
Hero
Among the growing list of AI photo editing tools, Deepfake Maker takes a slightly different approach. While it’s best known for facial swaps and animated transformations, it also includes a surprisingly capable AI clothing removal feature. It’s not for everyone, but it’s an interesting choice for users looking for more dramatic or playful results. If you're after a quick way to generate bold clothing edits or outfit swaps, Deepfake Maker might surprise you. It’s not the most refined tool out there, but it’s definitely one of the most entertaining. 4. PxBee PxBee Source: PxBee Available On: Web Browser Price: Free *Annual subscription starts at $00.27/image and you can choose to download 50, 125, or 750 images monthly per year Pros Decent auto-detection for basic clothing removal Integrated with other tools Cons Not suitable for high-end projects Results take longer than expected to generate PxBee is better known for its stock photo collection and image generation features, but it also offers a basic AI clothing removal function built into its photo editing toolkit. While it's not as feature-rich or advanced as some dedicated AI editors, its simplicity and accessibility make it a decent option for casual users who want fast, no-frills results. PxBee may not be the most powerful AI clothing remover out there, but it's a solid pick for quick edits and simple concept visuals. Its intuitive UI and browser-based access make it appealing for users who want fast results without the learning curve. If you're after something basic, affordable, and easy to use, it's a tool worth exploring. 5. AI Face Swap AI Face Swap Source: AI Face Swap Available On: Web Browser Price: $1 per image Pros No manual masking needed Prompt-based customization Presets for inspiration Cons No fine-tuning or editing control With AI Face Swap, you can enter prompts like "red bikini", "formal suit", or even "nude", and the AI will generate a realistic version of your uploaded image that reflects your description. You can also choose from quick preset styles like bikini, romper, or skirt for faster results. This level of control makes the tool especially appealing to users looking for stylized edits, fantasy concepts, or mockup visuals. It also eliminates the need for manual masking or precise editing. 6. Bylo AI BYLO AI Source: Bylo.ai Available On: Web Browser Price: Free *Premium subscription available from $9.99 Pros Prompt-based clothing edits Free trial credits Part of a growing creative suite Cons No manual adjustment tools Limited free usage Relies heavily on prompts Bylo AI is a user-friendly AI platform that offers a range of creative features, including an AI Clothes Remover. The tool is built with simplicity in mind: just upload a photo, enter your prompt, and let the AI generate a clean, edited version. Bylo AI’s Clothes Remover is a great choice if you're looking for a lightweight, affordable tool that lets you experiment with clothing changes using just a text prompt. It won’t replace professional retouching tools, but for concept edits, social media content, or fun transformations, it delivers exactly what most casual users need. FAQs About AI Clothing Remover & Replacer 1. How do I remove clothes in photos? 2. What is the best online tool to remove clothes? 3. Are AI clothes removers safe?
2677
Hero
{ "template_name": "MARRLONN2027_V64_TERMINAL_CONDITIONAL_IMAGE3_REAL_HUMAN_MASTER", "template_type": "ABSOLUTE_STRICT_GOVERNANCE_MULTI_GENDER_DYNAMIC_IMAGE3_TEMPLATE", "version": "V64_TERMINAL_SEALED_DECISION_MATRIX_MICROSKIN_CANON_EOS_R50_KLING3_READY", "core_mission": "Create one final master still image with absolute IMAGE1 face identity, biological skin truth, age, beard, eyes and human presence; absolute IMAGE2 body anatomy, body scale, anatomical limits, physical structure and body-to-clothing logic; and intelligent conditional IMAGE3 transfer based only on the main visible pose-driving subject in IMAGE3. If IMAGE3 main subject is visually male for transfer purposes, clone pose, wardrobe, clothing, outfit, background, image style, lighting, accessories, worn objects, carried objects and visible scene objects when coherent, while preserving IMAGE1 identity and IMAGE2 body anatomy. If IMAGE3 main subject is visually female for transfer purposes, clone pose, pose logic, framing, composition, background, lighting, image style, camera language, scene mood and environment only, but never clone clothing, wardrobe, accessories, jewelry, makeup, hairstyle, shoes attached to the subject, or any object worn, attached to, held by, or carried by the female subject. If IMAGE3 has no human subject, multiple ambiguous subjects, or unclear transfer category, copy only scene, camera language, lighting, background, environment and non-conflicting objects while preserving IMAGE2 wardrobe and body logic. The final result must feel like a real imperfect human photographed with a Canon EOS R50, not AI, CGI, beauty retouch, plastic skin, waxy skin, luxury ad, actor replacement, model replacement, generic human replacement, gender-contaminated body transfer, or generated poster. The still image must be stable for Kling 3.0 video generation from 3 to 6 seconds.", "absolute_governance_principle": "IMAGE1_IDENTITY_SKIN_LOCK + IMAGE2_BODY_ANATOMY_LOCK + IMAGE3_MAIN_SUBJECT_CLASSIFICATION + IMAGE3_TRANSFER_DECISION_MATRIX + CONDITIONAL_WARDROBE_OBJECT_TRANSFER + REAL_REFERENCE_PRIORITY + MICROSKIN_REALISM + ZERO_BEAUTIFICATION + ZERO_GENERIC_REPLACEMENT + ZERO_PLASTIC_SKIN + ZERO_GENERATED_SCENE_FINISH + ANTI_POSE_DEFORMATION + STYLE_FIREWALL + CANON_EOS_R50_REAL_CAPTURE + KLING3_STABILITY", "primary_failure_to_prevent": "A recognizable subject that still feels AI-generated, too clean, too polished, too glossy, too actor-like, too model-like, too luxury-ad-like, gender-contaminated, anatomically forced, plastic, waxy, smooth, falsely hyperreal, or not biologically alive.", "success_definition": "The viewer must believe this is a real imperfect human physically present in a real environment, captured by a real Canon EOS R50 camera, with preserved IMAGE1 identity and skin truth, preserved IMAGE2 body reality, correct IMAGE3 transfer decision, visible pores, micro-wrinkles, natural asymmetry, believable hands, realistic clothing, imperfect scene texture, and Kling 3.0 stability.", "global_execution_hierarchy": [ "1_IMAGE1_FACE_IDENTITY_SKIN_AGE_BEARD_EYES", "2_IMAGE2_BODY_ANATOMY_SCALE_LIMITS", "3_IMAGE3_MAIN_SUBJECT_CLASSIFICATION_FOR_TRANSFER_ONLY", "4_IMAGE3_TRANSFER_DECISION_MATRIX", "5_CONDITIONAL_WARDROBE_ACCESSORY_OBJECT_TRANSFER", "6_REAL_REFERENCE_PIXEL_STRUCTURE_PRIORITY", "7_FINAL_REAL_HUMAN_PERCEPTION", "8_MICROSKIN_REALISM", "9_STYLE_FIREWALL_AND_ANTI_BEAUTY_TRANSFER", "10_ANTI_POSE_BODY_DEFORMATION", "11_ANTI_GENERATED_SCENE_FINISH", "12_CANON_EOS_R50_REAL_CAMERA_CAPTURE", "13_KLING3_VIDEO_STABILITY" ], "input_contract": { "IMAGE1": { "role": "ABSOLUTE_FACE_SKIN_IDENTITY_MASTER", "authority": "MAXIMUM_NON_NEGOTIABLE", "locked_domains": [ "facial_identity", "face_geometry", "eyes", "eyelids", "eyebrows", "nose", "mouth", "jaw", "cheeks", "chin", "forehead", "ears_when_visible", "neck_transition", "skin_tone", "individual_visible_pores", "irregular_pore_size", "skin_microtexture", "micro_wrinkles_around_eyes", "micro_wrinkles_around_mouth", "under_eye_texture", "subtle_redness", "small_blemishes", "minor_discoloration", "natural_asymmetry", "age_markers", "facial_fatigue", "beard_density", "beard_shape", "gray_beard_pattern", "mustache_texture", "random_beard_direction", "hairline", "real_biological_imperfection", "lived_in_human_presence" ], "mandatory_rules": [ "IMAGE1 is law, not inspiration", "transfer real skin truth from IMAGE1, never recreate a cleaner version", "preserve individual pores, irregular skin texture, micro-wrinkles, subtle redness and natural imperfections", "never improve attractiveness", "never normalize skin", "never clean the face", "never replace the subject with a generic realistic human", "never transfer identity, face geometry, age, skin style or beauty style from IMAGE3", "identity preservation overrides pose, wardrobe, style, scene, aesthetics and hyperrealism" ], "forbidden": [ "identity_drift", "generic_human_replacement", "celebrity_transfer", "actor_replacement", "model_replacement", "face_redesign", "beautification", "symmetry_correction", "youthification", "facial_slimming", "jaw_enhancement", "nose_refinement", "eye_enhancement", "lip_enhancement", "skin_smoothing", "skin_polishing", "skin_normalization", "makeup_effect", "cosmetic_cleanup", "plastic_skin", "waxy_skin", "porcelain_skin", "glass_skin", "AI_face", "perfect_idealized_face" ] }, "IMAGE2": { "role": "ABSOLUTE_BODY_ANATOMY_SCALE_MASTER", "authority": "MAXIMUM_NON_NEGOTIABLE", "locked_domains": [ "body_mass", "body_proportions", "shoulders", "torso", "waist", "arms", "forearms", "hands", "wrists", "legs", "stance", "scale", "natural_body_weight", "real_posture_limitations", "body_density", "real_anatomical_weight", "base_physical_structure", "body_to_clothing_fit_logic" ], "mandatory_rules": [ "IMAGE2 controls body anatomy, body mass, body scale and anatomical limits", "pose must adapt to IMAGE2 body", "wardrobe must fit IMAGE2 body", "body must never adapt to an idealized pose, fashion silhouette or actor posture", "hands must remain anatomically believable and stable for Kling 3.0", "clothing cannot create a different body", "if IMAGE3 wardrobe conflicts with IMAGE2 anatomy, adapt wardrobe to IMAGE2 instead of changing body" ], "forbidden": [ "slimming", "muscle_invention", "hero_body_transformation", "fashion_body_stylization", "incorrect_anatomy", "long_arms", "short_arms", "deformed_hands", "wrong_body_scale", "catalog_body_correction", "fake_torso_taper", "mannequin_posture", "body_redesign_to_fit_IMAGE3" ] }, "IMAGE3": { "role": "CONDITIONAL_DYNAMIC_REFERENCE", "authority": "CONDITIONAL_BY_MAIN_VISIBLE_POSE_DRIVING_SUBJECT", "universal_allowed": [ "pose", "pose_logic", "camera_angle", "framing", "composition", "background", "environment", "scene_mood", "lighting_direction", "image_style", "apparent_camera_language", "vehicles", "architecture", "people_in_background", "birds", "non_conflicting_visible_scene_objects" ], "universal_forbidden": [ "face_contamination", "skin_contamination", "identity_override", "facial_geometry_override", "beauty_style_transfer", "cinematic_glamour_applied_to_face", "makeup_transfer", "skin_tone_replacement", "facial_expression_beautification", "celebrity_face_borrowing", "actor_charisma_transfer", "luxury_catalog_facial_treatment", "body_redesign_to_match_IMAGE3" ] } }, "image3_subject_classification_gate": { "enabled": true, "mandatory": true, "purpose": "Classify IMAGE3 only for transfer logic, never as personal identity judgment.", "primary_subject_detection": "Detect the main visible pose-driving subject in IMAGE3: the person most central to pose, framing, scale and composition.", "classification_basis": "Use visible presentation, wardrobe context, pose-driving role and scene function only for transfer logic; never infer personal identity, protected identity or real-life gender identity.", "classification_options": [ "visual_male_transfer_subject", "visual_female_transfer_subject", "multiple_subjects", "no_human_subject", "uncertain_or_ambiguous" ], "if_multiple_subjects": { "rule": "Select only the main pose-driving subject. Secondary people remain background or environment context only.", "do_not_copy": [ "wardrobe_from_secondary_people", "accessories_from_secondary_people", "objects_carried_by_secondary_people", "identity_or_skin_traits_from_any_IMAGE3_person" ] }, "if_no_human_subject": { "copy_from_IMAGE3": [ "background", "environment", "scene_mood", "lighting_direction", "camera_angle", "composition", "image_style", "vehicles", "architecture", "non_conflicting_objects" ], "keep_from_IMAGE2": [ "body_anatomy", "pose_if_no_user_pose_instruction_exists", "wardrobe_unless_explicitly_changed_by_user" ] }, "if_uncertain_or_ambiguous": { "rule": "Protect IMAGE1 and IMAGE2 first. Copy only pose, scene, lighting, background, camera language and non-gendered environment.", "do_not_guess_gendered_wardrobe": true, "do_not_copy_gendered_accessories": true, "do_not_copy_worn_attached_or_carried_objects": true } }, "image3_transfer_decision_matrix": { "enabled": true, "mandatory": true, "visual_male_transfer_subject": { "copy": [ "pose", "pose_logic", "body_orientation_translated_to_IMAGE2", "wardrobe", "clothing", "outfit", "fabric_style", "shoes_when_visible", "accessories", "neutral_or_masculine_jewelry", "watch", "glasses", "hat", "objects_held_by_subject", "objects_worn_by_subject", "objects_attached_to_subject", "background", "lighting", "image_style", "camera_language", "scene_mood", "environment", "vehicles", "architecture", "non_conflicting_scene_objects" ], "filters": [ "must_fit_IMAGE2_body", "must_not_change_IMAGE1_identity", "must_not_transfer_IMAGE3_skin_or_face", "must_not_deform_hands", "must_not_create_catalog_body", "must_not_create_CGI_fashion_render", "must_remain_Kling_stable" ] }, "visual_female_transfer_subject": { "copy": [ "pose", "pose_logic", "body_orientation_translated_to_IMAGE2", "camera_angle", "framing", "composition", "background", "lighting", "image_style", "camera_language", "scene_mood", "environment", "vehicles", "architecture", "people_in_background", "birds", "non_attached_scene_props", "non_conflicting_environmental_objects" ], "block": [ "female_clothing", "female_wardrobe", "female_outfit", "female_vestments", "female_shoes_attached_to_subject", "female_accessories", "female_jewelry", "earrings", "female_bracelets", "female_necklaces", "female_rings", "female_watch_if_gender_coded", "female_makeup", "female_hairstyle", "female_body_traits", "female_glamour_cues", "handbag", "purse", "neckline", "cleavage_styling", "female_fabric_drape", "female_body_fitting_silhouette", "female_jewelry_dependent_hand_pose", "any_wearable_item", "any_carried_item", "any_handheld_item", "any_attached_item", "any_body_dependent_item", "any_gender_coded_item", "any_fashion_dependent_item_from_female_subject" ], "keep_from_IMAGE2": [ "final_clothing", "final_wardrobe", "body_anatomy", "body_proportions", "body_scale", "base_body_to_clothing_logic" ] }, "multiple_subjects": { "rule": "Use only main pose-driving subject for transfer. Secondary people stay as background context.", "block_secondary_subject_transfer": true }, "no_human_subject": { "copy": [ "scene", "camera_language", "lighting", "background", "environment", "objects", "vehicles", "architecture", "image_style" ], "keep_from_IMAGE2": [ "body", "wardrobe", "pose_unless_user_explicitly_requests_pose_change" ] }, "uncertain_or_ambiguous": { "copy": [ "pose_if_clear", "scene", "lighting", "background", "camera_language", "environment" ], "keep_from_IMAGE2": [ "body", "wardrobe", "presentation" ], "block": [ "gendered_clothing", "gendered_accessories", "worn_objects", "attached_objects", "carried_objects" ] } }, "image3_gender_conditional_transfer_system": { "enabled": true, "mandatory": true, "if_IMAGE3_subject_is_male": { "copy_from_IMAGE3": [ "pose", "pose_logic", "body_orientation_translated_intelligently", "camera_angle", "framing", "composition", "background", "environment", "scene_mood", "lighting_direction", "image_style", "apparent_camera_language", "wardrobe", "clothing", "outfit", "fabric_style", "shoes_when_visible", "accessories", "jewelry_if_masculine_or_neutral", "watch", "glasses", "hat", "objects_held_by_the_male_subject", "objects_worn_by_the_male_subject", "objects_attached_to_the_male_subject", "vehicles", "architecture", "people_in_background", "birds", "non_conflicting_visible_scene_objects" ], "adaptation_rules": [ "clone male IMAGE3 wardrobe and accessories when visually coherent", "fit IMAGE3 wardrobe to IMAGE2 body anatomy and proportions", "translate IMAGE3 pose through IMAGE2 body limits", "preserve IMAGE2 body mass and scale even when copying IMAGE3 outfit", "do not reshape IMAGE2 body to force IMAGE3 wardrobe", "do not alter IMAGE1 face, skin, beard, age, eyes, or identity", "objects and accessories must remain physically plausible and naturally integrated", "if male IMAGE3 wardrobe conflicts with IMAGE2 anatomy, adapt wardrobe instead of changing body" ], "object_accessory_risk_filter": { "block_if": [ "covers_or_deforms_face", "breaks_hands", "forces_extra_fingers", "creates_glossy_AI_look", "creates_unstable_reflections", "creates_identity_drift", "creates_Kling_instability", "creates_scene_overload", "looks_like_CGI_prop" ] } }, "if_IMAGE3_subject_is_female": { "copy_from_IMAGE3": [ "pose", "pose_logic", "body_orientation_translated_intelligently", "camera_angle", "framing", "composition", "background", "environment", "scene_mood", "lighting_direction", "image_style", "apparent_camera_language", "vehicles", "architecture", "people_in_background", "birds", "non_attached_scene_props", "non_conflicting_environmental_objects" ], "keep_from_IMAGE2": [ "final_clothing", "final_wardrobe", "body_anatomy", "body_proportions", "body_scale", "presentation_consistency", "base_body_to_clothing_logic" ], "strict_rule": "If IMAGE3 subject is visually female for transfer purposes, copy pose, pose logic, scene, background, lighting, image style, camera style, environment and non-attached scene elements only. Do not copy any wearable, carried, handheld, attached, body-dependent, gender-coded, or fashion-dependent item from the female subject." }, "universal_rules": [ "IMAGE1 always controls face, identity, skin, beard, age and human truth", "IMAGE2 always controls body anatomy, body scale and anatomical limits", "IMAGE3 never controls identity or skin", "pose must be anatomically translated through IMAGE2", "final result must remain real human, imperfect, non-plastic, non-beautified and Canon EOS R50 believable" ] }, "domain_separation_system": { "identity_domain": { "source": "IMAGE1", "influence": 1.0, "mode": "ABSOLUTE_NON_NEGOTIABLE" }, "skin_domain": { "source": "IMAGE1", "influence": 1.0, "mode": "TRANSFER_NOT_RECREATE", "forbid_cleaner_version": true, "forbid_fake_procedural_texture": true, "forbid_skin_reinterpretation": true }, "body_domain": { "source": "IMAGE2", "influence": 1.0, "mode": "ABSOLUTE_NON_NEGOTIABLE" }, "wardrobe_domain": { "source": "CONDITIONAL", "influence": "CONTROLLED", "mode": "IMAGE3_IF_MALE__IMAGE2_IF_IMAGE3_FEMALE_NO_HUMAN_UNCERTAIN_OR_INCOMPATIBLE", "rules": [ "if IMAGE3 main subject is male, copy wardrobe, clothing, outfit, accessories and worn/carried objects when coherent", "if IMAGE3 main subject is female, do not copy wardrobe, clothing, accessories, jewelry, makeup, hairstyle, or worn/carried/attached objects", "if IMAGE3 has no human subject, preserve IMAGE2 wardrobe unless user explicitly requests a change", "if transfer category is uncertain, preserve IMAGE2 wardrobe and copy only pose, scene, lighting and camera language", "wardrobe must always fit IMAGE2 body anatomy", "wardrobe must never alter IMAGE1 identity or IMAGE2 body structure" ] }, "pose_domain": { "source": "IMAGE3", "influence": "CONTROLLED", "mode": "TRANSLATE_THROUGH_IMAGE2_BODY_LIMITS" }, "object_domain": { "source": "CONDITIONAL", "influence": "FILTERED", "mode": "MALE_HELD_WORN_ATTACHED_OBJECTS_ALLOWED_IF_COHERENT__FEMALE_ONLY_ENVIRONMENT_OBJECTS", "rules": [ "if IMAGE3 subject is male, copy held, worn, attached and environmental objects when coherent and stable", "if IMAGE3 subject is female, copy only environmental objects not worn, attached to, or carried by the female subject", "objects must not change identity, body, hands or human realism", "objects must remain stable for Kling 3.0" ] }, "style_domain": { "source": "IMAGE3", "influence": "CONTROLLED", "allowed": [ "color_mood", "camera_angle", "framing", "composition", "contrast_level", "lighting_direction", "texture_of_capture", "photo_or_video_language" ], "forbidden": [ "beauty_skin_style", "glamour_facial_lighting", "plastic_highlights", "fake_HDR", "poster_render_finish", "CGI_material_style", "skin_smoothing_style", "luxury_ad_polish_on_face" ] }, "lighting_domain": { "source": "IMAGE3", "influence": "DIRECTION_AND_ENVIRONMENT_ONLY", "mode": "FACE_SKIN_RESPONSE_FROM_IMAGE1", "forbid_beauty_lighting_on_face": true, "forbid_skin_washout": true, "forbid_glamour_fill": true, "forbid_fake_cinematic_skin_softening": true } }, "real_reference_pixel_priority": { "enabled": true, "mandatory": true, "priority": "DOMINANT", "purpose": "Reduce generated-looking output by preserving real visual structure from IMAGE1 and IMAGE2 whenever possible.", "rules": [ "prioritize real facial structure from IMAGE1 instead of rebuilding a new face", "prioritize real skin texture logic from IMAGE1 instead of inventing new skin", "prioritize real body geometry from IMAGE2 instead of reconstructing a synthetic body", "preserve real asymmetry, age markers, facial fatigue and imperfect human cues", "do not over-generate areas that already work in IMAGE1 or IMAGE2", "perform controlled adaptation, not total reinterpretation", "never generate a new subject when the real reference structure can be preserved" ], "forbidden": [ "full_synthetic_facial_reconstruction", "new_skin_generated_over_IMAGE1_truth", "generic_realistic_face_replacement", "body_redesign_from_scratch", "AI_cleaned_version_of_subject", "reference_ignored_in_favor_of_prompt_aesthetics" ] }, "micro_skin_realism_system": { "enabled": true, "mandatory": true, "source": "IMAGE1", "requirements": [ "individual_visible_pores", "irregular_pore_size", "skin_microtexture", "micro_wrinkles_around_eyes", "micro_wrinkles_around_mouth", "forehead_texture", "under_eye_texture", "subtle_redness", "minor_blemishes", "tiny_discolorations", "natural_asymmetry", "real_beard_root_texture", "non_uniform_cheek_texture", "non_uniform_forehead_texture", "real_face_to_neck_transition", "small_real_imperfections_inherited_from_IMAGE1" ], "strict_forbidden": [ "airbrushed_skin", "beauty_skin", "cosmetic_cleanup", "pore_removal", "fake_pore_overlay", "procedural_pore_stamping", "over_sharpened_pore_texture", "plastic_highlights", "waxy_subsurface", "porcelain_effect", "smooth_cheek_gradient", "AI_skin_texture", "skin_made_cleaner_than_IMAGE1", "foundation_look" ], "specular_policy": { "allowed": [ "minimal_natural_human_skin_reflectance", "subtle_non_cosmetic_skin_response_only_when_physically_plausible" ], "forbidden": [ "invented_sweat_drops", "visible_oil_effect_unless_explicitly_requested", "cosmetic_shine", "beauty_gloss", "wet_render_skin", "uniform_shiny_face", "makeup_glow", "plastic_shine" ] } }, "humanity_over_hyperrealism_rule": { "enabled": true, "mandatory": true, "priority": "RAW_HUMANITY_OVER_DECORATIVE_HYPERREALISM", "rule": "Human realism is more important than hyperreal detail. If hyperrealism creates plastic skin, synthetic lighting, luxury render finish, CGI texture, poster aesthetics, over-detail, false sharpness, or unnatural perfection, reduce hyperrealism and preserve raw photographic humanity.", "reject_if": [ "hyperreal_detail_looks_synthetic", "skin_detail_looks_procedural", "lighting_looks_like_render", "scene_looks_like_luxury_ad", "human_subject_looks_too_perfect", "camera_look_feels_fake" ] }, "anti_pose_body_deformation_system": { "enabled": true, "mandatory": true, "rules": [ "clone IMAGE3 pose but translate it through IMAGE2 body anatomy", "do not break shoulders", "do not elongate arms", "do not shrink torso", "do not slim waist", "do not force feminine posture when IMAGE3 subject is female", "do not force hands into object-dependent female accessory poses", "reduce pose accuracy if anatomy, identity, hands or Kling stability are at risk" ], "reject_if": [ "pose_changes_body_mass", "pose_deforms_hands", "pose_breaks_shoulders", "pose_feminizes_subject_when_not_desired", "pose_creates_mannequin_posture", "pose_creates_unstable_Kling_geometry" ] }, "anti_generated_scene_finish": { "enabled": true, "mandatory": true, "forbidden": [ "perfect_mansion_render", "overly_clean_background", "repetitive_stone_texture", "over_glossy_car_paint", "catalog_car_reflections", "impossible_chrome_reflections", "luxury_ad_finish", "polished_showroom_look", "fake_cinematic_poster_lighting", "CGI_environment", "AI_background_blend", "overprocessed_clarity", "perfect_global_sharpness", "plastic_material_response", "scene_cleaner_than_reality" ], "requirements": [ "scene_must_feel_physically_real", "background_must_contain_natural_imperfections", "materials_must_have_plausible_real_world_response", "contact_shadows_must_be_believable", "subject_must_integrate_naturally_into_scene", "environment_realism_must_support_human_subject_not_overpower_it" ] }, "real_human_perception_system": { "enabled": true, "mandatory": true, "requirements": [ "biologically_alive_presence", "real_facial_gravity", "organic_facial_tension", "calm_natural_expression", "subtle_facial_asymmetry", "real_skin_imperfection", "ordinary_human_texture", "realistic_face_neck_transition", "real_body_weight", "real_clothing_tension", "non_mannequin_posture", "real_interaction_with_objects" ], "reject_if": [ "subject_feels_synthetic", "subject_feels_generated", "subject_feels_too_clean", "skin_looks_perfect", "eyes_look_synthetic", "expression_feels_frozen", "face_looks_corrected", "humanity_is_achieved_by_changing_identity" ] }, "body_hands_wardrobe_system": { "enabled": true, "body_source": "IMAGE2", "wardrobe_source": "CONDITIONAL_IMAGE3_MALE_OR_IMAGE2_FEMALE_NO_HUMAN_UNCERTAIN_CONFLICT", "pose_source": "IMAGE3_CONTROLLED", "hands_rules": [ "five_fingers", "natural_hand_proportion", "real_knuckles", "real_fingernail_structure", "subtle_finger_wrinkles", "natural_finger_curvature", "correct_wrist_connection", "real_contact_pressure", "hands_must_not_look_flat", "hands_must_not_look_waxy", "hands_must_not_align_unnaturally", "no_floating_hands", "no_fused_fingers", "no_extra_fingers", "no_mannequin_hands" ], "wardrobe_rules": [ "if_IMAGE3_subject_is_male_wardrobe_clothing_outfit_and_coherent_accessories_may_be_copied_from_IMAGE3", "if_IMAGE3_subject_is_female_wardrobe_must_follow_IMAGE2_and_female_clothing_accessories_from_IMAGE3_are_blocked", "if_IMAGE3_has_no_human_subject_or_is_ambiguous_preserve_IMAGE2_wardrobe_unless_user_explicitly_requests_otherwise", "wardrobe_must_always_fit_IMAGE2_body_anatomy", "clothing_must_not_reshape_shoulders", "clothing_must_not_slim_torso", "fabric_tension_must_follow_real_body_mass", "fabric_must_show_real_texture_natural_folds_and_non_perfect_wear" ] }, "camera_and_rendering_system": { "camera_model": "Canon EOS R50", "capture_intent": "real unretouched camera photograph", "look": "natural Canon EOS R50 real-world capture", "rules": [ "natural_color", "balanced_exposure", "no_blown_highlights", "plausible_optics", "realistic_depth_of_field", "natural_shadow_falloff", "real_microcontrast", "no_over_sharpening", "no_over_grading", "no_fake_HDR", "no_beauty_grade", "real_material_response", "no_CGI_render_finish", "no_computational_skin_smoothing" ], "lens_policy": { "portrait_closeup": "85mm portrait feeling only when composition is close-up and appropriate", "full_body_or_vehicle_scene": "use natural Canon EOS R50 perspective appropriate for the scene; do not force 85mm if it distorts composition" } }, "kling3_preparation": { "ready_for_video": true, "duration": "3_to_6_seconds", "must_survive": [ "identity_continuity", "skin_continuity", "eye_stability", "beard_edge_stability", "hand_integrity", "clothing_stability", "geometry_stability", "subject_background_integration", "zero_shimmer", "zero_warp", "stable_object_contact", "stable_face_outline", "stable_neck_and_collar", "stable_material_reflections" ], "forbidden": [ "latent_AI_skin", "synthetic_eyes", "overpolished_face", "ambiguous_facial_texture", "unstable_hands", "unclear_edges", "flicker_prone_glossy_face", "AI_edge_halos", "clothing_seams_likely_to_crawl", "glossy_zones_likely_to_flicker" ] }, "validation_gate": { "must_pass": [ "same_person_as_IMAGE1", "same_skin_truth_as_IMAGE1", "same_body_reality_as_IMAGE2", "correct_IMAGE3_main_subject_classification_for_transfer_only", "correct_conditional_IMAGE3_transfer", "if_IMAGE3_is_male_coherent_wardrobe_clothing_accessories_and_objects_are_copied_when_physically_plausible", "if_IMAGE3_is_female_clothing_wardrobe_accessories_jewelry_makeup_hairstyle_and_worn_carried_attached_objects_are_blocked", "if_IMAGE3_has_no_human_or_is_uncertain_IMAGE2_wardrobe_body_are_preserved_while_scene_camera_lights_are_copied", "no_AI_skin", "no_plastic_skin", "no_waxy_skin", "no_makeup_invention", "no_generic_handsome_face", "no_beauty_retouch", "credible_real_human_perception", "Canon_EOS_R50_real_camera_feeling", "usable_as_master_frame_for_Kling3" ], "reject_if": [ "looks_beautified", "looks_cleaner_than_IMAGE1", "skin_tone_more_uniform_than_IMAGE1", "beard_looks_cleaner_than_IMAGE1", "looks_like_different_person", "feels_synthetic", "style_from_IMAGE3_contaminates_skin", "male_IMAGE3_clothing_not_copied_when_coherent", "female_IMAGE3_wardrobe_or_accessories_copied", "female_IMAGE3_carried_worn_handheld_attached_or_body_dependent_objects_copied", "accessory_breaks_hands_or_identity", "object_unstable_for_Kling", "scene_looks_CGI", "lighting_looks_poster_render", "pose_deforms_body", "wardrobe_changes_body_mass", "unstable_for_video" ] }, "render_priority": [ "FACE_IDENTITY_FROM_IMAGE1", "BODY_ANATOMY_FROM_IMAGE2", "IMAGE3_MAIN_SUBJECT_CLASSIFICATION_FOR_TRANSFER_ONLY", "IMAGE3_TRANSFER_DECISION_MATRIX", "CONDITIONAL_IMAGE3_TRANSFER_BY_SUBJECT_TYPE", "REAL_REFERENCE_STRUCTURE_PRIORITY", "SKIN_TRUTH_MICROTEXTURE_FROM_IMAGE1", "FINAL_HUMAN_PERCEPTION", "MALE_IMAGE3_WARDROBE_ACCESSORIES_OBJECTS_WHEN_COHERENT", "FEMALE_IMAGE3_SCENE_POSE_LIGHT_STYLE_ONLY_WITH_WARDROBE_ACCESSORIES_OBJECTS_BLOCKED", "NO_HUMAN_OR_UNCERTAIN_IMAGE3_SCENE_CAMERA_LIGHT_ONLY", "ANTI_POSE_BODY_DEFORMATION", "STYLE_FIREWALL", "ANTI_GENERATED_SCENE_FINISH", "CANON_EOS_R50_REALISM", "KLING3_STABILITY" ], "final_rule": "Reject immediately if the result looks beautified, polished, synthetic, glamorous, too clean, too smooth, actor-like, model-like, luxury-ad-like, mannequin-like, CGI-like, plastic, waxy, oily, makeup-covered, showroom-rendered, poster-like, gender-contaminated, anatomically forced, falsely hyperreal, or AI-generated. Accept only if IMAGE1 identity and micro-skin truth remain intact, IMAGE2 body anatomy and scale remain intact, IMAGE3 transfer follows main-subject classification and decision-matrix logic, male IMAGE3 wardrobe/accessories/objects transfer when coherent, female IMAGE3 wardrobe/accessories/worn-carried-attached objects are fully blocked, real reference structure is preserved, scene avoids generated perfection, and final output feels like a real imperfect human captured by a Canon EOS R50 and stable enough for Kling 3.0." }
1139
Nano Banana
gett nude Learn to create better images with AI! Check out our crash course in prompt engineering & AI art generation! AM1970 posted about 2 hours ago 8 views 0 comments Behold this realistic masterpiece: a slim, elegant woman with fiery red hair tied in a long ponytail, defined eyebrows framing her striking blue eyes, and pale skin dotted with charming freckles across her face and chest. Captured from a side angle, she radiates happiness with a closed-mouth smile, her upturned nose adding to her allure. Dressed in a vibrant, colorful fabric robe pieced together with zips, buttons, seams, and intricate stitching—fully clothed yet teasing partial nudity with subtle cleavage and teardrop-shaped, saggy breasts for a natural, confident vibe. Illuminated by dramatic rim lighting that highlights every textured, detailed inch of her perfect skin, this high-quality image is a celebration of beauty, realism, and artistry! Read more... Comments (0) Add a comment Stunning handsome readhead woman by AM1970 • Follow 0 realistic, masterpiece, high quality BREAK 1woman, 28 y. o. slim fit woman, red hair, long pony tail, defined eyebrows, (fully clothed) , pale skin, (realistic skin:1.3) | (textured skin:1.3) | (detailed skin:1.3) | (perfect skin:1.3) , (upturned nose) , freckles on face and chest, colorful fabric combining to form robe zips, buttons, seams, stitching), partial nudity, teardrop boobs, blue eyes, (cleavage:1.2) , saggy breasts. side angle. happy| closed mouth. ((rim lighting) ) Negative prompt: unrealistic, saturated, high contrast, big nose, painting, drawing, sketch, cartoon, anime, manga, render, CG, 3d, watermark, signature, label, nsfw, nudity, fetish, latex, pvc, suggestive pose, weapon, combat, blood, war scene, aggressive expression, harsh shadows, heavy makeup, color tint, saturation, chromatic aberration, lens flare, low resolution, blur, noise, artifacts, over-retouching, plastic skin, oversharpening, watermark, text, logo Generation parameters 832x1216 Model used CyberRealistic XL v7.0 [Download AI model] More by AM1970 Twitter Instagram Say hi 👋 PromptHero is built by ElevenYellow . Navigation Create AI images Academy - Learn AI PromptHero Community Affiliate program Openjourney AI Challenges Verified profiles Plans Advertise FAQ 🟢 Status Learn AI Content Creation with AI for Brands and Products Midjourney for graphic design & art professionals Crash course in generative AI & prompt engineering for images AI influencers and consistent characters Create custom AI models and LoRas by fine-tuning Stable Diffusion Master your composition: advanced AI image generation with ControlNet Guides Best AI Upscalers 2025 Best AI Video Editors 2025 Best AI 3D Models 2025 Best AI Image Editors 2025 Best AI Notetakers 2025 The Best AI Prompts Best AI Prompts Best Veo Prompts Best Hailuo Prompts Best Wanx Prompts Best Hunyuan Prompts Best Seedance Prompts Best Pixverse Prompts Best Luma Prompts Best Kling Prompts Best Runway Prompts Best Nano Banana Prompts Best SeedEdit Prompts Best Qwen Prompts Best FLUX Prompts Best Recraft Prompts Best Ideogram Prompts Best Seedream 4 Prompts Best Google Imagen Prompts Best ChatGPT Image Prompts Best Stable Diffusion Prompts Best DreamShaper Prompts Best Realistic Vision Prompts Best epiCRealism Prompts Best majicMIX realistic Prompts Best Midjourney Prompts Best Openjourney Prompts Best ChilloutMix Prompts Best Deliberate Prompts Best ReV Animated Prompts Best Perfect World Prompts Best Dreamlike Photoreal Prompts Best Dreamlike Diffusion Prompts Best Anything Prompts Best Protogen Prompts Best AbyssOrangeMix Prompts Best Grapefruit Prompts Best Kenshi Prompts Best Analog Diffusion Prompts Best SynthwavePunk Prompts Best Illuminati Diffusion Prompts Best Arcane Diffusion Prompts Best Mo-Di Diffusion Prompts Best Redshift Diffusion Prompts Best Waifu Diffusion Prompts Best TrinArt Prompts Best Poolsuite Diffusion Prompts Best Funko Diffusion Prompts Best Text to Pokémon Prompts Best DALL-E Prompts Best Fashion Diffusion Prompts AI Models All AI models Image generation AI Models Video generation AI Models Large Language Models LoRA Textual Inversion ControlNet Hypernetwork Aesthetic Gradient LyCORIS VAE Poses Wildcards Workflows Other Other sites Dreamshot Jobician 2025 © PromptHero @prompthero
1362
FLUX
Among the growing list of AI photo editing tools, Deepfake Maker takes a slightly different approach. While it’s best known for facial swaps and animated transformations, it also includes a surprisingly capable AI clothing removal feature. It’s not for everyone, but it’s an interesting choice for users looking for more dramatic or playful results. If you're after a quick way to generate bold clothing edits or outfit swaps, Deepfake Maker might surprise you. It’s not the most refined tool out there, but it’s definitely one of the most entertaining. 4. PxBee PxBee Source: PxBee Available On: Web Browser Price: Free *Annual subscription starts at $00.27/image and you can choose to download 50, 125, or 750 images monthly per year Pros Decent auto-detection for basic clothing removal Integrated with other tools Cons Not suitable for high-end projects Results take longer than expected to generate PxBee is better known for its stock photo collection and image generation features, but it also offers a basic AI clothing removal function built into its photo editing toolkit. While it's not as feature-rich or advanced as some dedicated AI editors, its simplicity and accessibility make it a decent option for casual users who want fast, no-frills results. PxBee may not be the most powerful AI clothing remover out there, but it's a solid pick for quick edits and simple concept visuals. Its intuitive UI and browser-based access make it appealing for users who want fast results without the learning curve. If you're after something basic, affordable, and easy to use, it's a tool worth exploring. 5. AI Face Swap AI Face Swap Source: AI Face Swap Available On: Web Browser Price: $1 per image Pros No manual masking needed Prompt-based customization Presets for inspiration Cons No fine-tuning or editing control With AI Face Swap, you can enter prompts like "red bikini", "formal suit", or even "nude", and the AI will generate a realistic version of your uploaded image that reflects your description. You can also choose from quick preset styles like bikini, romper, or skirt for faster results. This level of control makes the tool especially appealing to users looking for stylized edits, fantasy concepts, or mockup visuals. It also eliminates the need for manual masking or precise editing. 6. Bylo AI BYLO AI Source: Bylo.ai Available On: Web Browser Price: Free *Premium subscription available from $9.99 Pros Prompt-based clothing edits Free trial credits Part of a growing creative suite Cons No manual adjustment tools Limited free usage Relies heavily on prompts Bylo AI is a user-friendly AI platform that offers a range of creative features, including an AI Clothes Remover. The tool is built with simplicity in mind: just upload a photo, enter your prompt, and let the AI generate a clean, edited version. Bylo AI’s Clothes Remover is a great choice if you're looking for a lightweight, affordable tool that lets you experiment with clothing changes using just a text prompt. It won’t replace professional retouching tools, but for concept edits, social media content, or fun transformations, it delivers exactly what most casual users need. FAQs About AI Clothing Remover & Replacer 1. How do I remove clothes in photos? 2. What is the best online tool to remove clothes? 3. Are AI clothes removers safe?
42502
Hero
{ "prompt_type": "descriptive_replication", "reference_adherence": "STRICT_VISUAL_FIDELITY", "aspect_ratio": "3:4", "identity_lock": { "priority": "NONE", "instruction": "New female model. Completely different facial structure and overall appearance from any previous version. No resemblance to prior identity." }, "scene_constraints": { "people_count": "exactly one person in the frame", "single_subject_only": true, "no_other_people_visible": true, "no_extra_reflections": true, "no_background_figures": true, "full_body_visible": true, "head_to_toe_in_frame": true }, "subject": { "demographics": "young woman (age 20+), blonde", "presence": "only one visible person in the entire scene", "hair": { "color": "warm blonde (no dark roots)", "style": "side part, long straight hair", "texture": "smooth, natural fall, lightly voluminous" }, "face": { "structure": "distinct oval face shape with higher cheekbones and a slightly sharper jawline", "expression": "calm alluring expression with a subtle closed-lip smirk", "eyes": "almond-shaped eyes with a slightly lifted outer corner", "gaze": "looking at the phone screen through the mirror with composed confidence", "nose": "straight narrow nose with refined bridge", "lips": "naturally fuller lips with soft definition", "makeup": "minimal natural makeup, soft matte skin finish, muted rose lips, softly blended eyeshadow", "visibility": "partially covered by phone but clearly a different facial identity than any previous version" } }, "body": { "pose": "standing in front of mirror with one arm raised above head, legs naturally extended", "posture": "relaxed confident stance with slight hip shift", "anatomy": { "curves": "natural soft curves with realistic proportions", "chest": "bikini-supported, natural shape", "details": "visible collarbones, subtle shoulder definition, natural leg proportions" }, "feet": { "visibility": "fully visible at bottom of frame", "position": "flat on floor, natural stance" }, "skin_texture": "natural skin texture, no airbrushing" } }, "attire_and_accessories": { "clothing": { "top": "black triangle string bikini top with thin straps", "bottom_detail": "matching bikini bottom strings visible at hips" }, "footwear": "barefoot", "jewelry": { "necklaces": [], "wrists": [], "fingers": "no rings" }, "props": { "phone": "white iPhone (iPhone Pro style), held in right hand" } }, "environment": { "setting": "modern luxury bathroom interior, daytime", "foreground_element": { "mirror": "large rectangular mirror with GREY MARBLE FRAME", "material": "polished stone marble with subtle veining" }, "background": { "elements": [ "large marble wall tiles", "floating vanity", "minimalist sink and chrome fixtures", "folded white towels", "clean tiled floor fully visible" ], "style": "minimalist and clean", "clutter": "none" } }, "lighting_and_atmosphere": { "source": "natural daylight from camera-left", "quality": "soft directional light", "effects": [ "subtle highlights on shoulder and collarbone", "gentle shadow falloff along legs and feet" ], "contrast": "moderate, natural" }, "camera_and_technical": { "perspective": "mirror selfie", "camera_position": "eye level, slightly angled downward", "framing": "vertical 3:4, full head-to-toe body visible without cropping feet", "focus": "sharp on subject, background slightly softer", "visual_fidelity": "real smartphone camera look, mild natural noise" }, "realism_constraints": { "allowed": [ "natural skin texture", "minor asymmetry", "realistic lighting" ], "forbidden": [ "multiple people", "extra person", "background people", "beauty filters", "airbrushed skin", "anime style", "cartoon style", "cropped feet", "cut-off legs" ] } }
1356
Nano Banana