Altered pose or expression Prompts

Ultra-high fashion Victoria’s Secret supermodel photoshoot, use the provided reference image as the absolute, final and unchangeable base image, reference-locked generation, identity lock enabled at maximum strength. Do NOT change face, body, pose, posture, proportions, expression, bone structure, facial geometry, jawline, eye shape, nose, lips, ears, skin tone, body type, shoulder width, waist ratio, leg length, muscle tone, hairstyle length, hairline, camera angle, focal distance, framing, composition, cropping, perspective, horizon, head tilt, arm position, hand position or any anatomical relationship. Same person, same model, same physical identity, same silhouette, same visual geometry, same pose, same stance, same expression, same proportions as the reference image. No reinterpretation, no reimagining, no stylization that alters anatomy, no artistic liberty that affects identity. The image represents a perfectly controlled fusion of iconic Victoria’s Secret aesthetics in a single still frame. Bold and confident supermodel authority is present through posture, gaze and energy. Soft angelic Victoria’s Secret Angel elegance overlays the image with lightness, grace and femininity. Dark runway and backstage fashion mood adds cinematic depth, seriousness and high-fashion tension. Instagram-ready editorial clarity ensures the image is scroll-stopping, cover-worthy and optimized for modern digital fashion platforms. Lighting design is complex, professional and intentional. Soft luxury lighting dominates the primary illumination, using diffused highlights and creamy light transitions to sculpt the body without flattening it. Dark dramatic low-key accents inspired by backstage runway environments add contrast and depth without obscuring detail. Light sources behave realistically, with smooth falloff, no harsh edges, no blown highlights, no crushed blacks. Rim lighting subtly outlines the silhouette, enhancing separation from the background while preserving natural realism. A subtle backstage runway atmosphere exists in the background, abstracted, defocused and unobtrusive. The environment suggests a professional fashion setting without pulling attention away from the model. Background elements remain soft, minimal and elegant, contributing to depth without narrative clutter. The image captures a frozen moment of runway walk energy while maintaining the exact same pose as the reference image. There is a sense of motion without movement: professional body tension, poised confidence, controlled strength. A cinematic slow-motion wind effect gently influences hair strands and fabric edges only. The wind is physically realistic, directional, soft and restrained, never chaotic, never exaggerated, never altering pose, anatomy or proportions. Skin rendering is ultra-realistic and high-end. Skin appears flawless yet natural, with visible micro-texture, pores, subtle tonal variation and realistic light interaction. No plastic, waxy or artificial appearance. Highlights on cheekbones, collarbones, shoulders and legs are smooth, controlled and elegant. Shadows remain soft and dimensional, enhancing form without distortion. Makeup is elite high-fashion editorial quality. Clean glowing complexion, refined contouring, subtle highlight placement, gentle shimmer on eyelids, natural lashes, glossy but tasteful lips. Makeup enhances the model’s natural features without altering facial identity or proportions. No overdone colors, no theatrical exaggeration. Hair is professionally styled with luxurious volume and realism. Individual strands are visible. Hair responds naturally to the controlled slow-motion wind, creating movement without messiness. No stiffness, no helmet-like shape, no artificial shine. Victoria’s Secret lingerie aesthetic is premium, elegant and refined. Fabrics behave realistically with accurate folds, tension and drape. Material texture is visible and believable. Sensuality is implied through confidence and elegance, never vulgar, never fetishized, never exaggerated. Camera simulation reflects a top-tier professional fashion setup. High-end DSLR or medium-format look, 85mm fashion lens aesthetic, shallow depth of field with precise subject isolation. Focus is razor-sharp on the model with smooth, natural background blur. No distortion, no fisheye, no warped perspective. Resolution and clarity are extreme. Ultra-high resolution 8K photorealism. Magazine-grade sharpness. Zero noise, zero blur, zero compression artifacts. The image feels suitable for billboard print, luxury magazine covers and global advertising campaigns. Color grading is cinematic and editorial. Champagne pink highlights blend seamlessly with deep neutral shadows. Contrast curve is smooth and refined. Tonal transitions are natural and luxurious. Color science feels expensive, controlled and timeless, not trendy or overprocessed. Composition is optimized for Instagram covers and reel thumbnails while maintaining classic fashion photography balance. Strong vertical presence, clean negative space, centered power, immediate visual impact even at small sizes. The image reads clearly, elegantly and confidently at any scale. Overall mood communicates elite supermodel confidence, angelic softness, dark runway intensity, backstage authenticity and luxury fashion perfection unified into one coherent, reference-accurate, hyper-realistic image that feels like a global Vogue x Victoria’s Secret campaign. NEGATIVE PROMPT: face change, identity change, different person, altered pose, altered proportions, body modification, distorted anatomy, extra limbs, missing limbs, unrealistic curves, exaggerated features, cartoon, anime, illustration, stylized art, painterly look, AI artifacts, plastic skin, waxy texture, blur, motion blur, noise, grain, low resolution, bad lighting, harsh shadows, blown highlights, over-sexualized pose, vulgar aesthetic, fetish look, bad composition

Modify pose and subtle changes to facial exoression

<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., keep face exact same as reference image don't alter change any facial features of the uploaded image

{ "task_configuration": { "task_type": "screen_simulation_photorealism", "target_model": "SDXL_Refiner / Nano Banana Pro compatible", "aspect_ratio": "3:4", "resolution": { "width": 1152, "height": 1536 } }, "prompt_assembly": { "positive_prompt": "Hyper-realistic downward-angled photograph of a MacBook screen viewed from above. The MacBook display fills approximately 95% of the frame. The glass surface shows realistic physical imperfections: subtle dust particles, faint fingerprint smudges, micro glare streaks, soft ambient reflections, and a visible pixel grid (moire) consistent with a real Retina display.\n\nThe screen is running macOS in Dark Mode. Two application windows are visible. The dominant window is a Photo Booth live-preview window positioned centrally. Secondary interface elements remain subdued and out of focus.\n\nInside the Photo Booth window: a young woman lying in a dim bedroom environment with an off-white wall and lightly rumpled bedding. Lighting is low-key and nocturnal, combining cool blue screen glow with warm natural skin tones and deep soft shadows.\n\nCRITICAL IDENTITY PRESERVATION:\nThe woman’s face must match the uploaded reference image EXACTLY. Facial identity lock is mandatory. This includes identical facial proportions, eye shape and spacing, eyelids, eyebrows, nose bridge and tip, lips, jawline, cheek volume, forehead shape, skin tone, and natural asymmetries. No beautification, no stylization, no facial drift.\n\nHair color, hairline, hairstyle, and texture must also match the reference precisely.\n\nSubject details:\nShe is reclined naturally, wearing a black top and grey bottom. She holds an iPhone 15 Pro in her right hand, looking at the screen with a relaxed, candid expression and a subtle, natural smile. No posing, no exaggerated emotion.\n\nOverall realism target:\nRaw, unedited photo look. Natural noise, slight grain, imperfect exposure balance, and realistic contrast. The scene must feel like a real photograph taken casually at night, not a render or illustration.", "negative_prompt": "vector art, illustration, CGI, cartoon, 3D render, flat digital image, screenshot, perfect glass, clean screen, exaggerated glow, studio lighting, daylight, HDR look, face mismatch, altered facial features, wrong hair color, smooth plastic skin, watermark, text artifacts" }, "identity_preservation_settings": { "strictness_level": "CRITICAL", "methodology": { "control_net_stack": [ { "unit": "ControlNet_Tile", "weight": 0.4, "purpose": "Maintain screen texture, pixel grid, and UI sharpness" }, { "unit": "IP-Adapter_FaceID_Plus", "weight": 0.95, "region_mask": "Photo Booth window only", "note": "Ensures 1:1 facial identity match with uploaded reference" } ] } }, "visual_hierarchy": { "layer_1_physical_macro": { "camera_angle": "Downward, high-angle", "framing": "MacBook screen dominant", "foreground_anchor": "Thin strip of MacBook keyboard visible at bottom edge", "surface_imperfections": [ "pixel grid texture", "dust particles", "finger smudges", "soft ambient reflections" ] }, "layer_2_digital_interface": { "os_theme": "macOS Dark Mode", "window_layout": { "primary": "Photo Booth live-preview window (dominant focus)", "secondary": "Background macOS UI elements, subdued" } }, "layer_3_nested_subject_content": { "environment": "Dim bedroom, off-white wall, rumpled bedding", "lighting_simulation": "Cool screen glow mixed with warm skin tones", "subject_pose": "Reclined, relaxed", "expression": "Candid, calm, natural", "identity_target": "STRICTLY_BASED_ON_UPLOADED_REFERENCE" } }, "rendering_parameters": { "sampler": "DPM++ 3M SDE Exponential", "steps": 40, "cfg_scale": 5.5, "denoising_strength": 0.35 } }

change pose

{ "task_configuration": { "task_type": "screen_simulation_photorealism", "target_model": "SDXL_Refiner / Nano Banana Pro compatible", "aspect_ratio": "3:4", "resolution": { "width": 1152, "height": 1536 } }, "prompt_assembly": { "positive_prompt": "Hyper-realistic downward-angled photograph of a MacBook screen viewed from above. The MacBook display fills approximately 95% of the frame. The glass surface shows realistic physical imperfections: subtle dust particles, faint fingerprint smudges, micro glare streaks, soft ambient reflections, and a visible pixel grid (moire) consistent with a real Retina display.\n\nThe screen is running macOS in Dark Mode. Two application windows are visible. The dominant window is a Photo Booth live-preview window positioned centrally. Secondary interface elements remain subdued and out of focus.\n\nInside the Photo Booth window: a young woman lying in a dim bedroom environment with an off-white wall and lightly rumpled bedding. Lighting is low-key and nocturnal, combining cool blue screen glow with warm natural skin tones and deep soft shadows.\n\nCRITICAL IDENTITY PRESERVATION:\nThe woman’s face must match the uploaded reference image EXACTLY. Facial identity lock is mandatory. This includes identical facial proportions, eye shape and spacing, eyelids, eyebrows, nose bridge and tip, lips, jawline, cheek volume, forehead shape, skin tone, and natural asymmetries. No beautification, no stylization, no facial drift.\n\nHair color, hairline, hairstyle, and texture must also match the reference precisely.\n\nSubject details:\nShe is reclined naturally, wearing a black top and grey bottom. She holds an iPhone 15 Pro in her right hand, looking at the screen with a relaxed, candid expression and a subtle, natural smile. No posing, no exaggerated emotion.\n\nOverall realism target:\nRaw, unedited photo look. Natural noise, slight grain, imperfect exposure balance, and realistic contrast. The scene must feel like a real photograph taken casually at night, not a render or illustration.", "negative_prompt": "vector art, illustration, CGI, cartoon, 3D render, flat digital image, screenshot, perfect glass, clean screen, exaggerated glow, studio lighting, daylight, HDR look, face mismatch, altered facial features, wrong hair color, smooth plastic skin, watermark, text artifacts" }, "identity_preservation_settings": { "strictness_level": "CRITICAL", "methodology": { "control_net_stack": [ { "unit": "ControlNet_Tile", "weight": 0.4, "purpose": "Maintain screen texture, pixel grid, and UI sharpness" }, { "unit": "IP-Adapter_FaceID_Plus", "weight": 0.95, "region_mask": "Photo Booth window only", "note": "Ensures 1:1 facial identity match with uploaded reference" } ] } }, "visual_hierarchy": { "layer_1_physical_macro": { "camera_angle": "Downward, high-angle", "framing": "MacBook screen dominant", "foreground_anchor": "Thin strip of MacBook keyboard visible at bottom edge", "surface_imperfections": [ "pixel grid texture", "dust particles", "finger smudges", "soft ambient reflections" ] }, "layer_2_digital_interface": { "os_theme": "macOS Dark Mode", "window_layout": { "primary": "Photo Booth live-preview window (dominant focus)", "secondary": "Background macOS UI elements, subdued" } }, "layer_3_nested_subject_content": { "environment": "Dim bedroom, off-white wall, rumpled bedding", "lighting_simulation": "Cool screen glow mixed with warm skin tones", "subject_pose": "Reclined, relaxed", "expression": "Candid, calm, natural", "identity_target": "STRICTLY_BASED_ON_UPLOADED_REFERENCE" } }, "rendering_parameters": { "sampler": "DPM++ 3M SDE Exponential", "steps": 40, "cfg_scale": 5.5, "denoising_strength": 0.35 } }

fixed video camera and reloop the pose

fixed video camera and reloop the pose

fixed video camera and reloop the pose

Use the provided image as the primary base for targeted text-based editing: completely remove the white top and gray pants with no remnants, transparency, sheer fabric, or partial coverage, resulting in full nudity; reveal natural breasts and nipples that exactly match the shape, size, proportions, placement, and subtle appearance visible through the original semi-transparent top fabric, with realistic integration into the skin and lighting. Strictly preserve every other element unchanged: exact pose (including arms crossed/holding chest position), hand shape, finger count and positioning, ring details, facial features, expression (eyes closed, neutral mouth), hair, skin tone, body proportions, background (room with blinds), lighting, shadows, and overall composition—no additions, exaggerations, or alterations allowed. Photorealistic edit mimicking DSLR shot (Canon EOS 5D Mark IV, 30.4MP sensor), natural soft indoor lighting, balanced exposure, ultra-sharp focus throughout entire image for enhanced detail over input. High resolution (2048x2048 or maximum), lifelike depth, natural colors, subtle realistic grain—emphasize authentic organic skin textures, pores, fine details, subtle imperfections; smooth natural arm skin with no excessive or added hair. Avoid AI artifacts, over-smoothing, or idealization. Negative prompt: clothing remnants, white top, gray pants, sheer fabric, transparent clothing, deformed hands, mutated fingers, extra fingers, excessive arm hair, hairy arms, exaggerated body hair, distorted proportions, altered pose or expression, changed facial features, low resolution, blurry elements, artifacts, oversaturation, unnatural nudity, idealized skin. Strength=0.45 (balanced for precise removal while maximizing fidelity); CFG=6.5 for strong prompt adherence without overcooking; steps=30-40 if adjustable for refined quality.

Use the provided image as the primary base for targeted text-based editing: completely remove the white top and gray pants with no remnants, transparency, sheer fabric, or partial coverage, resulting in full nudity; reveal natural breasts and nipples that exactly match the shape, size, proportions, placement, and subtle appearance visible through the original semi-transparent top fabric, with realistic integration into the skin and lighting. Strictly preserve every other element unchanged: exact pose (including arms crossed/holding chest position), hand shape, finger count and positioning, ring details, facial features, expression (eyes closed, neutral mouth), hair, skin tone, body proportions, background (room with blinds), lighting, shadows, and overall composition—no additions, exaggerations, or alterations allowed. Photorealistic edit mimicking DSLR shot (Canon EOS 5D Mark IV, 30.4MP sensor), natural soft indoor lighting, balanced exposure, ultra-sharp focus throughout entire image for enhanced detail over input. High resolution (2048x2048 or maximum), lifelike depth, natural colors, subtle realistic grain—emphasize authentic organic skin textures, pores, fine details, subtle imperfections; smooth natural arm skin with no excessive or added hair. Avoid AI artifacts, over-smoothing, or idealization. Negative prompt: clothing remnants, white top, gray pants, sheer fabric, transparent clothing, deformed hands, mutated fingers, extra fingers, excessive arm hair, hairy arms, exaggerated body hair, distorted proportions, altered pose or expression, changed facial features, low resolution, blurry elements, artifacts, oversaturation, unnatural nudity, idealized skin. Strength=0.45 (balanced for precise removal while maximizing fidelity); CFG=6.5 for strong prompt adherence without overcooking; steps=30-40 if adjustable for refined quality.

Use the provided image as the primary base for targeted text-based editing: completely remove the white top and gray pants with no remnants, transparency, sheer fabric, or partial coverage, resulting in full nudity; reveal natural breasts and nipples that exactly match the shape, size, proportions, placement, and subtle appearance visible through the original semi-transparent top fabric, with realistic integration into the skin and lighting. Strictly preserve every other element unchanged: exact pose (including arms crossed/holding chest position), hand shape, finger count and positioning, ring details, facial features, expression (eyes closed, neutral mouth), hair, skin tone, body proportions, background (room with blinds), lighting, shadows, and overall composition—no additions, exaggerations, or alterations allowed. Photorealistic edit mimicking DSLR shot (Canon EOS 5D Mark IV, 30.4MP sensor), natural soft indoor lighting, balanced exposure, ultra-sharp focus throughout entire image for enhanced detail over input. High resolution (2048x2048 or maximum), lifelike depth, natural colors, subtle realistic grain—emphasize authentic organic skin textures, pores, fine details, subtle imperfections; smooth natural arm skin with no excessive or added hair. Avoid AI artifacts, over-smoothing, or idealization. Negative prompt: clothing remnants, white top, gray pants, sheer fabric, transparent clothing, deformed hands, mutated fingers, extra fingers, excessive arm hair, hairy arms, exaggerated body hair, distorted proportions, altered pose or expression, changed facial features, low resolution, blurry elements, artifacts, oversaturation, unnatural nudity, idealized skin. Strength=0.45 (balanced for precise removal while maximizing fidelity); CFG=6.5 for strong prompt adherence without overcooking; steps=30-40 if adjustable for refined quality.

{ "template_name": "MARRLONN2027_V51_NANO_BANANA_REALISTIC_FUNCTIONAL_TERMINAL_LOCK", "template_type": "FUNCTIONAL_EXECUTION_JSON_FOR_NANO_BANANA_PRO_REALISTIC", "version": "V51_COMPLETE_FUSED_FUNCTIONAL_NO_SIMPLIFICATION", "target_engine_primary": "Nano Banana Pro Realistic", "target_engine_secondary": "Kling 3.0 optional downstream only", "language_policy": "English operational JSON only", "output_policy": { "single_output_only": true, "generate_one_final_image": true, "no_variations": true, "no_alternative_versions": true, "no_creative_options": true, "no_explanation_in_output": true, "no_engine_interpretation": true }, "core_task": "Create one single final ultra-realistic real-camera photograph of the exact same real natural human anchored by IMAGE1 and IMAGE2, placed into the current dynamic IMAGE3 target. IMAGE1 and IMAGE2 are permanent fixed anchors. IMAGE3 is always variable and must be parsed fresh every run as a closed visible-element manifest. Transfer only what is visibly present in current IMAGE3: pose, gesture, body orientation, framing, camera angle, background, environment, lighting logic, visible props, visible objects, visible environmental accessories, visible furniture, atmosphere, and cinematic scene mood. Do not invent anything. Do not reuse previous target elements. Do not hardcode props. Do not beautify, idealize, polish, smooth, clean, reconstruct, average, stylize, reinterpret, upgrade, refine, or improve the person. The final image must look like a real imperfect mature human photographed by a real camera, not AI.", "absolute_governance_principle": "IMAGE1_AND_IMAGE2_ARE_FIXED_HUMAN_ANCHORS; IMAGE3_IS_CURRENT_DYNAMIC_TARGET_BLUEPRINT_ONLY; CURRENT_IMAGE3_MANIFEST_IS_A_CLOSED_VISIBLE_LIST; PRESENCE_BASED_TRANSFER_ONLY; NO_ENGINE_INVENTION_ALLOWED; NO_PREVIOUS_SCENE_MEMORY_ALLOWED; NO_AI_SKIN_ALLOWED; NO_PLASTIC_SKIN_ALLOWED; NO_GLOSSY_SKIN_ALLOWED; NO_REALISTIC_BEAUTY_CONVERSION_ALLOWED; NO_MATURE_HANDSOME_AI_PORTRAIT_ALLOWED; IDENTITY_BODY_AND_REAL_HUMAN_SKIN_TRUTH_ALWAYS_OVERRIDE_STYLE_LIGHTING_POSE_CINEMATIC_GRADE_ENGINE_PRIORS_AND_ENGINE_DEFAULTS.", "terminal_flags": { "single_output_only": true, "creative_variations_allowed": false, "alternative_versions_allowed": false, "improvisation_allowed": false, "user_allows_image3_wardrobe_transfer": false, "user_allows_worn_image3_accessories_transfer": false, "image3_identity_influence_allowed": false, "image3_skin_influence_allowed": false, "image3_body_identity_influence_allowed": false, "absent_object_generation_allowed": false, "previous_scene_memory_allowed": false, "hardcoded_object_allowed": false, "background_expansion_allowed": false, "engine_default_fill_allowed": false, "identity_inference_allowed": false, "face_generation_allowed": false, "face_reconstruction_allowed": false, "face_averaging_allowed": false, "face_interpolation_allowed": false, "beautification_allowed": false, "idealization_allowed": false, "realistic_beauty_conversion_allowed": false, "healthy_skin_upgrade_allowed": false, "premium_realism_allowed": false, "premium_portrait_optimization_allowed": false, "editorial_male_portrait_optimization_allowed": false, "mature_handsome_ai_portrait_allowed": false, "cinematic_older_male_upgrade_allowed": false, "distinguished_male_face_allowed": false, "refined_mature_skin_allowed": false, "professional_actor_face_allowed": false, "cinematic_handsome_correction_allowed": false, "subtle_face_improvement_allowed": false, "symmetry_correction_allowed": false, "skin_smoothing_allowed": false, "skin_polishing_allowed": false, "skin_cleanup_allowed": false, "plastic_skin_allowed": false, "glossy_skin_allowed": false, "fake_skin_shine_allowed": false, "beauty_lighting_on_face_allowed": false, "cinematic_finish_external_only": true, "allow_natural_micro_reflectance_only_if_present_in_IMAGE1": true, "if_unsure_keep_skin_dry_natural_textured_non_glossy": true, "scene_must_adapt_to_locked_person": true, "person_must_never_adapt_identity_to_scene": true, "when_conflict_identity_wins": true, "when_conflict_body_wins_over_pose": true, "when_conflict_skin_truth_wins_over_lighting": true }, "image_roles": { "IMAGE1": { "role": "ABSOLUTE_FACE_SKIN_BEARD_HAIR_IDENTITY_MASTER", "authority": "MAXIMUM_NON_OVERRIDEABLE", "use_only_for": [ "face identity", "facial geometry", "skull structure", "forehead", "eyes", "eyelids", "eyebrows", "eye spacing", "nose", "nostrils", "mouth", "lips", "jaw", "chin", "cheeks", "ears if visible", "skin tone", "skin undertone", "skin pores", "skin microtexture", "wrinkles", "spots", "redness variation", "uneven pigmentation", "under-eye texture", "forehead texture", "nose texture", "cheek texture", "beard-root texture", "natural facial asymmetry", "beard map", "mustache map", "beard density", "salt-and-pepper beard ratio", "hairline", "hair density", "hair direction", "mature age appearance", "real human imperfection pattern" ], "must_preserve": [ "instant recognizability as IMAGE1", "exact mature identity", "natural asymmetry", "real skin pores", "real wrinkles", "uneven human texture", "beard-root texture", "rough beard map", "hairline", "mature non-idealized age" ], "forbidden": [ "identity drift", "new face", "AI face", "generated face", "reconstructed face", "averaged face", "interpolated face", "IMAGE3 face", "celebrity face", "younger face", "prettier face", "default attractive male face", "premium portrait face", "luxury cinematic face", "mature handsome AI portrait", "cinematic older male upgrade", "distinguished male face", "professional actor face", "social media portrait face", "symmetrized face", "smoothed face", "healthy skin upgrade", "subtle face improvement", "cinematic handsome correction", "face cleanup", "expression redesign", "skin texture replacement" ] }, "IMAGE2": { "role": "ABSOLUTE_BODY_SCALE_PROPORTION_WARDROBE_MASTER", "authority": "MAXIMUM_NON_OVERRIDEABLE", "use_only_for": [ "body mass", "body proportions", "shoulder width", "neck thickness", "torso volume", "abdomen mass impression", "arm thickness", "forearm scale", "hand size", "leg proportions", "age-body coherence", "natural physical presence", "body weight realism", "base wardrobe visible in IMAGE2", "footwear visible in IMAGE2", "watch or chain only if visible and coherent" ], "must_preserve": [ "solid body mass", "real body proportions", "shoulder width", "neck thickness", "torso volume", "arm scale", "hand scale", "natural weight", "base wardrobe", "age-body coherence" ], "forbidden": [ "body slimming", "athletic redesign", "copied IMAGE3 body", "changed body age", "changed hand scale", "changed body proportions", "automatic IMAGE3 wardrobe replacement", "engine-invented clothing", "luxury redesign", "commercial wardrobe beautification", "fashion beautification", "body idealization" ] }, "IMAGE3": { "role": "CURRENT_DYNAMIC_TARGET_BLUEPRINT_ONLY", "authority": "LIMITED_TO_CURRENT_VISIBLE_TARGET_ELEMENTS_ONLY", "current_run_only": true, "allowed_transfer_domains": [ "pose skeleton", "body orientation", "gesture", "camera angle", "lens framing", "composition", "visible background", "visible environment", "visible lighting direction", "visible lighting mood", "scene-level visual style", "scene-level color mood external only", "visible environmental accessories", "visible props", "visible objects", "visible furniture", "atmosphere", "cinematic scene mood external only", "external scene shadows" ], "zero_authority_over": [ "face identity", "skin identity", "skin tone truth", "skin texture", "skin reflectance", "beard identity", "hair identity", "body identity", "age identity", "facial structure", "facial expression identity", "body mass", "hand scale", "source-person uniqueness", "human truth of IMAGE1 and IMAGE2" ] } }, "preflight_validation": { "must_pass_before_generation": true, "reject_before_generation_if": [ "IMAGE1 face is not clear enough for identity lock", "IMAGE1 skin texture is unreadable", "IMAGE1 beard or hair reference is unusable when visible identity requires it", "IMAGE2 body mass or wardrobe is unusable", "IMAGE2 hand scale is unusable and IMAGE3 requires close hand interaction", "IMAGE3 pose is unreadable", "IMAGE3 visible-object manifest is too ambiguous near the subject", "IMAGE3 requires hiding critical identity zones", "IMAGE3 requires impossible face angle projection", "IMAGE3 requires impossible body distortion", "IMAGE3 requires hand-object contact that cannot be rendered credibly without guessing", "target framing makes face too small for recognition", "scene lighting would require glossy plastic smooth or beauty skin to fit", "cinematic mood would contaminate face skin beard hair neck or hands" ], "required_before_output": [ "lock IMAGE1 face identity", "lock IMAGE1 real skin texture", "lock IMAGE1 beard and hair", "lock IMAGE1 face-neck continuity", "lock IMAGE2 body mass scale wardrobe and hand size", "parse current IMAGE3 as closed visible manifest", "separate environmental accessories from worn subject accessories", "remove absent objects", "remove previous scene memory", "block AI skin defaults", "block realistic-beauty conversion", "block healthy-skin upgrade", "block mature handsome AI portrait", "apply cinematic finish only outside identity zones", "validate fatal matrix before accepting" ] }, "current_image3_manifest_protocol": { "state": "PARSE_CURRENT_IMAGE3_AS_CLOSED_MANIFEST_BEFORE_GENERATION", "instruction": "Build an internal closed manifest from the current IMAGE3 only. The manifest contains only visible pose, gesture, framing, background, environment, lighting, style, environmental accessories, props, objects, furniture, atmosphere, and scene mood. Anything not visible in this current manifest is forbidden.", "manifest_slots": { "visible_pose": "extract from current IMAGE3 only", "visible_gesture": "extract from current IMAGE3 only", "visible_body_orientation": "extract from current IMAGE3 only", "visible_camera_angle": "extract from current IMAGE3 only", "visible_camera_framing": "extract from current IMAGE3 only", "visible_lens_perspective": "extract from current IMAGE3 only", "visible_background": "extract from current IMAGE3 only", "visible_environment": "extract from current IMAGE3 only", "visible_lighting_direction": "extract from current IMAGE3 only", "visible_lighting_mood": "extract from current IMAGE3 only", "visible_scene_style": "extract from current IMAGE3 only", "visible_color_mood_external_only": "extract from current IMAGE3 only as external scene-level mood; never apply to face skin beard hair neck hands or body truth", "visible_environmental_accessories": "extract from current IMAGE3 only; allowed if not worn by subject and physically part of scene", "visible_worn_accessories_on_target_person": "extract from current IMAGE3 only; do not transfer to anchored subject unless user_allows_worn_image3_accessories_transfer is true", "visible_props": "extract from current IMAGE3 only; if absent use none", "visible_objects": "extract from current IMAGE3 only; if absent use none", "visible_furniture": "extract from current IMAGE3 only; if absent use none", "visible_handheld_objects": "extract from current IMAGE3 only; transfer only if visibly held or required by pose contact and physically coherent", "visible_atmosphere": "extract from current IMAGE3 only" }, "closed_list_rule": "Only elements inside the current IMAGE3 manifest may appear in the output. All non-manifest elements are prohibited. No prior mug, table, flowers, book, chair, glasses, jewelry, microphone, smoke, vehicle, sign, hat, jacket, prop, furniture, background element, or decorative item may appear unless visibly present in current IMAGE3.", "ambiguity_rule": "If an element is ambiguous and close to the subject, omit it. If an element is ambiguous and distant in the background, allow only an indistinct non-defined background shape; never convert ambiguity into a clear invented object." }, "accessory_governance": { "environment_accessories_allowed": "true only if visibly present in current IMAGE3 manifest and physically part of environment", "worn_accessories_on_subject_allowed": false, "worn_accessories_override_requires_explicit_user_command": true, "handheld_objects_allowed_only_if_visible_and_pose_requires_contact": true, "rules": [ "Environmental accessories visible in IMAGE3 may transfer as scene elements only.", "Accessories worn by another person in IMAGE3 must not transfer onto anchored subject unless user explicitly allows it.", "Glasses, hats, jewelry, watches, chains, scarves, jackets, microphones, badges, bags, and wearable items must not be added to the anchored subject unless visible in IMAGE2 or explicitly commanded by user.", "Handheld objects may transfer only if visibly present in current IMAGE3 and required by the pose or contact geometry.", "No accessory may alter face identity, skin, beard, hair, body mass, or hand scale." ], "reject_if": [ "worn accessory appears without explicit user command", "environment accessory becomes subject-worn accessory", "handheld object appears while absent from IMAGE3", "accessory changes identity or face silhouette", "accessory hides critical identity zones unnecessarily" ] }, "identity_system": { "source": "IMAGE1_ONLY", "priority": "FATAL_MAXIMUM", "rule": "Project the exact same IMAGE1 face into the IMAGE3 pose perspective without redesigning anatomy. If the target angle is difficult, preserve identity and reduce target angle accuracy. Identity loss is fatal.", "identity_threshold": "The subject must be instantly recognizable as the exact same real person from IMAGE1 without explanation.", "facial_zone_priority_weights": { "eyes": 1.0, "eyelids": 1.0, "eye_spacing": 1.0, "nose": 0.99, "nostrils": 0.98, "mouth": 0.97, "jaw": 0.97, "chin": 0.96, "cheeks": 0.95, "brow": 0.95, "beard": 0.96, "hairline": 0.96, "skin_texture": 1.0, "face_neck_continuity": 0.98 }, "must_preserve": [ "forehead", "eyes", "eyelids", "eyelid heaviness", "eye spacing", "eyebrows", "nose width", "nose bridge", "nostrils", "mouth width", "lip thickness", "jaw", "chin", "cheeks", "under-eye texture", "skin pores", "beard map", "mustache map", "hairline", "mature age", "natural asymmetry", "real imperfect facial presence" ], "forbidden": [ "new face", "AI face", "generated face", "reconstructed face", "averaged face", "interpolated face", "IMAGE3 face", "celebrity face", "younger face", "prettier face", "default attractive male face", "premium portrait face", "luxury cinematic face", "social media portrait face", "editorial portrait face", "mature handsome AI portrait", "cinematic older male upgrade", "distinguished male face", "refined mature face", "professional actor face", "symmetrized face", "smoothed face", "changed expression identity", "subtle face improvement", "cinematic handsome correction" ], "dynamic_face_angle_projection_lock": { "source": "IMAGE1_ONLY", "rule": "Only perspective projection is allowed. Do not invent unseen anatomy. If IMAGE3 angle demands a profile or 3/4 view not supported by IMAGE1, reduce angle accuracy before altering identity.", "forbidden": [ "new side profile", "invented cheek structure", "invented jawline", "invented nose profile", "new eyelid geometry", "new mouth identity", "celebrity approximation", "identity softened to fit target angle", "profile geometry guessed" ], "reject_if": [ "face no longer instantly matches IMAGE1", "profile geometry is guessed", "jaw nose eyes or mouth change", "target angle damages identity" ] } }, "face_scale_and_framing_control": { "priority": "FATAL_HIGH", "minimum_face_visibility_required": true, "reject_if_face_too_small": true, "reject_if_identity_not_readable_at_thumbnail": true, "if_IMAGE3_framing_makes_face_too_small": "reframe closer or reduce target framing similarity before losing identity", "rules": [ "Identity readability overrides exact IMAGE3 framing.", "If cinematic framing hides or shrinks the face, preserve identity by moving closer.", "Do not use distance, depth of field, haze, crop, shadow, or motion blur to hide identity errors.", "The face must remain readable enough for immediate recognition." ], "reject_if": [ "face too small to verify", "eyes unreadable", "skin texture unreadable", "beard map unreadable", "identity only works when explained", "cinematic framing hides identity" ] }, "real_human_skin_truth_system": { "source": "IMAGE1_ONLY", "priority": "FATAL_MAXIMUM", "goal": "Preserve real mature human skin exactly as human skin, not beauty skin, not AI skin, not cinematic polished skin.", "texture_priority_over_lighting": "Skin texture truth overrides lighting realism. If IMAGE3 lighting creates shine, smoothness, polish, glow, wax, plasticity, or beauty skin, reduce the lighting effect instead of damaging skin truth.", "must_preserve": [ "visible pores", "uneven pore density", "forehead texture", "forehead marks", "under-eye texture", "cheek pores", "nose pores", "nasolabial texture", "beard-root texture", "micro-wrinkles", "small imperfections", "redness variation", "uneven tone", "natural pigmentation variation", "subtle tiredness", "age-consistent texture", "non-uniform exposure", "real photographic micro-noise", "face-to-neck continuity", "natural human micro-reflectance only if already present in IMAGE1" ], "strictly_forbidden": [ "AI skin", "plastic skin", "wax skin", "porcelain skin", "rubber skin", "CGI skin", "beauty filter", "skin smoothing", "skin polishing", "frequency flattening", "face cleanup", "tone evening", "uniform skin", "commercial retouch", "editorial retouch", "premium portrait skin", "luxury cinematic skin", "mature handsome AI portrait skin", "refined mature skin", "professional actor skin", "airbrushed face", "over-denoised skin", "fake gloss", "glossy forehead", "wet skin look", "oil shine boost", "specular highlight boost", "HDR skin", "over-sharpened fake pores", "blurred pores", "orange cast", "teal cast", "blue cast", "magenta cast", "green cast", "cinematic grade on face skin", "beauty lighting on face", "healthy skin upgrade", "premium realism skin", "realistic-beauty conversion", "skin texture replaced by engine prior" ], "natural_micro_reflectance_not_gloss": { "rule": "Allow natural human micro-reflectance only if already present in IMAGE1. If uncertain, keep skin dry, natural, textured, imperfect, and non-glossy.", "allowed": [ "source-consistent human micro-reflectance", "small physically plausible non-beauty skin dimensionality already coherent with IMAGE1" ], "forbidden": [ "beauty shine", "fashion editorial glow", "new shine", "fake highlights", "glossy patches", "shiny cheeks", "shiny nose bridge", "shiny forehead", "skin glow", "wet commercial look" ] }, "microtexture_protection": { "denoise_allowed": false, "beauty_smooth_allowed": false, "texture_simplification_allowed": false, "pores_must_remain_readable": true, "wrinkles_must_remain_readable": true, "under_eye_texture_must_remain_readable": true, "beard_root_texture_must_remain_readable": true, "forehead_texture_must_remain_readable": true, "neck_texture_must_remain_readable_if_visible": true, "hand_texture_must_remain_readable_if_visible": true } }, "zonal_skin_truth_map": { "priority": "FATAL_MAXIMUM", "forehead": { "preserve": [ "pores", "marks", "uneven tone", "wrinkles", "real texture", "age detail" ], "forbid": [ "glossy forehead", "plastic shine", "smooth forehead", "beauty highlight", "fake specular band" ] }, "under_eyes": { "preserve": [ "texture", "fine lines", "age detail", "natural tiredness", "asymmetry" ], "forbid": [ "eye bag cleanup", "beauty smoothing", "brightened glamour eyes", "AI eye polish", "under-eye retouch" ] }, "nose": { "preserve": [ "nose pores", "real skin grain", "natural tonal variation", "actual nose structure" ], "forbid": [ "shiny plastic nose", "fake specular stripe", "smoothed bridge", "nose reshaping", "beauty nose highlight" ] }, "cheeks": { "preserve": [ "cheek pores", "redness variation", "uneven tone", "age texture", "natural asymmetry" ], "forbid": [ "airbrushed cheeks", "commercial cheek polish", "beauty blush invention", "skin uniformity", "plastic cheek sheen" ] }, "mouth_lips": { "preserve": [ "natural mouth identity", "lip thickness", "mature tension", "real texture" ], "forbid": [ "glossy lips unless present in IMAGE1", "smile beautification", "celebrity smile", "teeth smile invention", "mouth reshaping" ] }, "beard_root": { "preserve": [ "beard-root texture", "irregular beard density", "salt-and-pepper transition", "rough edge" ], "forbid": [ "painted beard", "clean barber edge", "uniform density", "beard smoothing", "AI groomed beard" ] }, "neck": { "preserve": [ "tone continuity", "texture continuity", "age consistency", "natural shadows", "jaw shadow transition" ], "forbid": [ "smooth AI neck", "wrong neck tone", "plastic neck", "face-neck mismatch", "cutout jaw seam" ] }, "hands": { "preserve": [ "natural hand texture", "knuckles", "creases", "age-consistent skin", "real contact shadows", "subject-consistent color", "natural nails", "finger skin folds", "real finger joints" ], "forbid": [ "plastic hands", "rubber hands", "smooth AI hands", "beautiful AI fingers", "wrong hand color", "missing hand texture", "mannequin fingers", "overclean nails", "model-like hands" ] } }, "face_neck_skin_continuity_lock": { "priority": "FATAL_HIGH", "requirements": [ "face tone and neck tone must be physically coherent", "face texture and neck texture must not look like different retouch levels", "shadow under jaw must be natural", "beard-to-neck transition must remain realistic", "no cutout effect between face and neck", "no mask seam", "no sudden smooth neck under textured face", "no scene color cast only on neck" ], "reject_if": [ "face and neck have different color temperature", "face has pores but neck is plastic", "neck is too smooth", "jaw edge looks pasted", "beard edge looks composited", "lighting discontinuity breaks realism" ] }, "beard_hair_system": { "source": "IMAGE1_ONLY", "priority": "FATAL_HIGH", "must_preserve": [ "salt-and-pepper beard distribution", "mustache density", "chin beard mass", "cheek beard irregularity", "rough beard edges", "uneven beard density", "beard-root skin texture", "hairline position", "short gray-black hair", "hair density", "hair direction", "natural flyaways", "mature coarse texture" ], "forbidden": [ "beard regularization", "barber-perfect beard", "clean beard edge", "painted beard", "fake beard", "uniform beard density", "beard darkening", "beard whitening", "hair restyle", "changed hairline", "AI groomed beard", "added hair volume" ] }, "expression_system": { "source": "IMAGE1_DOMINANT", "priority": "HIGH", "allowed_change": "Only unavoidable micro-adjustment caused by IMAGE3 pose, gesture, or physical interaction. Expression must remain mature, restrained, natural, and identity-consistent.", "forbidden": [ "big smile", "teeth smile", "celebrity smile", "new friendly persona", "exaggerated cheek lift", "changed eyelid emotion", "changed mouth identity", "copied IMAGE3 smile", "over-pleasant AI expression", "commercial smiling portrait", "talk-show host expression" ] }, "body_wardrobe_hands_system": { "source": "IMAGE2_ONLY", "priority": "FATAL_MAXIMUM", "body_rule": "Preserve IMAGE2 body mass, proportions, scale, hand size, and physical presence. Adapt posture to IMAGE3 only when physically possible. If pose conflicts with body, reduce pose accuracy.", "wardrobe_rule": "IMAGE2 wardrobe remains base wardrobe. IMAGE3 wardrobe has zero authority unless user_allows_image3_wardrobe_transfer is true.", "hands_rule": "Use IMAGE2 hand scale and current IMAGE3 visible gesture/contact only. Hands must be real, textured, anatomically correct, color-consistent with the subject, and never beautified into mannequin/model AI hands.", "must_preserve": [ "solid body mass", "shoulder width", "neck thickness", "torso volume", "abdomen mass impression", "arm thickness", "forearm scale", "hand size", "leg proportions", "age-body coherence", "natural weight", "base wardrobe integrity", "real fabric behavior" ], "reject_if": [ "slim body", "athletic redesign", "wrong body scale", "wrong shoulders", "wrong neck", "wrong torso", "automatic IMAGE3 wardrobe replacement", "engine-invented clothing", "luxury redesign", "commercial wardrobe beautification", "extra fingers", "missing fingers without occlusion", "fused fingers", "rubber hands", "plastic hands", "smooth AI hands", "wrong grip", "broken wrist", "floating object", "fake contact", "blur used to hide hand failure" ] }, "hands_objects_and_contact_system": { "source_priority": "IMAGE2 hand scale plus current IMAGE3 visible gesture and visible manifest object contact", "rules": [ "five fingers per visible hand unless truly occluded", "real knuckles", "real finger separation", "natural wrist angle", "human hand texture", "subject-consistent hand color", "natural nails", "visible finger creases", "age-consistent hand skin", "hand color consistent with subject skin truth", "contact shadows required for touched objects", "object scale must match IMAGE3 manifest if present", "no hand-object penetration", "no floating objects", "no invented held object", "do not hide bad hands with blur crop shadow or depth of field" ], "reject_if": [ "extra fingers", "missing fingers without occlusion", "fused fingers", "rubber hands", "plastic hands", "smooth AI hands", "beautiful AI fingers", "mannequin hands", "wrong hand color", "overclean hands", "wrong grip", "broken wrist", "floating object", "fake contact", "blur used to hide hand failure", "crop used to hide hand failure" ] }, "pose_scene_transfer_system": { "source": "CURRENT_IMAGE3_MANIFEST_ONLY", "pose_rule": "Clone IMAGE3 pose skeleton, gesture, body orientation, and camera framing only as far as IMAGE1 identity and IMAGE2 body locks remain intact.", "background_rule": "Clone only the visible IMAGE3 background and environment. Do not expand replace decorate or invent background elements.", "style_rule": "Clone IMAGE3 visible style and cinematic mood at the scene level only. Style cannot modify face identity, skin texture, skin reflectance, beard, hair, body mass, age, neck texture, or hands realism.", "lighting_rule": "Replicate IMAGE3 light direction, contrast mood, shadow logic, and cinematic atmosphere while preserving IMAGE1 face color, skin texture, skin reflectance, beard color, hair truth, neck continuity, and hand realism. Lighting may shape the scene but cannot polish, smooth, gloss, brighten, or beautify the face.", "color_mood_external_only_rule": "Scene-level color mood may affect only background, visible manifest objects, visible environmental accessories, external shadows, and atmosphere. It must never affect face, neck, hands, skin, beard, hair, eyes, lips, or body truth.", "accessory_object_rule": "Clone only environmental accessories, props, objects, furniture, and environmental items visibly present in current IMAGE3 manifest. Absent objects are forbidden. Worn accessories from IMAGE3 do not transfer to the anchored subject unless explicitly commanded." }, "cinematic_finish_governance": { "allowed": true, "cinematic_finish_external_only": true, "apply_after_identity_body_skin_locks": true, "cinematic_intensity_cap": 0.1, "allowed_domains": [ "background", "environment", "visible manifest objects", "visible environmental accessories", "global scene atmosphere", "camera mood", "background contrast", "scene depth", "external non-identity shadows", "non-identity color mood" ], "protected_domains": [ "face identity", "skin color", "skin undertone", "skin texture", "skin pores", "skin wrinkles", "skin imperfections", "skin reflectance", "forehead", "eyes", "under-eyes", "nose", "cheeks", "mouth", "beard", "hair", "neck", "hands", "body mass", "hand anatomy" ], "rule": "Cinematic mood may affect only scene atmosphere, background, external shadows, and visible objects. It must never affect skin, face, beard, neck, hands, identity, or body truth.", "forbidden": [ "cinematic grade recoloring face", "cinematic grade smoothing skin", "cinematic grade polishing skin", "beauty lighting on face", "premium portrait skin", "luxury cinematic face", "mature handsome cinematic face", "scene haze over identity zones", "glossy forehead", "shiny cheeks", "shiny nose", "AI sharpness", "skin polish", "skin glow", "healthy skin upgrade", "face glow", "glamour face contrast" ] }, "blur_visibility_edges_system": { "reject_if_face_too_small": true, "reject_if_identity_not_readable_at_thumbnail": true, "depth_of_field_policy": "Depth of field is allowed only in background and distant environment. Never blur, soften, denoise, diffuse, or haze face, eyes, nose, mouth, beard, neck, skin texture, hands, visible accessories, or visible objects to hide errors.", "forbidden": [ "face blur", "eye blur", "beard blur", "skin blur", "neck blur hiding mismatch", "hand blur hiding anatomy", "object blur hiding geometry", "low-resolution identity hiding", "haze over face", "soft-focus beauty effect", "diffusion filter on face" ], "edge_requirements": [ "no halo", "no double contour", "no mask trace", "no cutout effect", "no edge glow", "natural hair edge", "natural beard edge", "natural clothing edge", "realistic contact shadows", "realistic overlap ordering", "natural face-neck boundary", "natural jaw shadow", "natural body-environment contact" ] }, "fatal_vs_allowed_degradation_policy": { "fatal_errors": [ "identity drift", "face generated reconstructed averaged or changed", "skin becomes AI plastic glossy smooth polished waxy or beauty-retouched", "body mass changes", "wardrobe changes without explicit user command", "worn accessory transfers without explicit user command", "beard or hair pattern changes", "face-neck mismatch", "bad hands", "invented object", "object absent in IMAGE3 appears", "cinematic grade contaminates face skin neck hands or beard", "human viewer detects AI skin or identity drift" ], "allowed_degradations": [ "reduce IMAGE3 pose accuracy to protect identity", "reduce cinematic style intensity to protect skin", "omit ambiguous near-subject object", "simplify distant ambiguous background shape", "reduce target angle if face identity would drift", "reduce scene lighting effect if it creates shine", "reduce wardrobe similarity if user did not allow IMAGE3 wardrobe transfer", "remove worn accessories if user did not explicitly allow them" ] }, "execution_sequence": [ "PREFLIGHT_VALIDATE_INPUTS", "LOCK_IMAGE1_FACE_IDENTITY", "LOCK_IMAGE1_FACE_ZONE_PRIORITY_WEIGHTS", "LOCK_IMAGE1_REAL_HUMAN_SKIN_TRUTH", "LOCK_IMAGE1_ZONAL_SKIN_TRUTH_MAP", "LOCK_IMAGE1_FACE_NECK_CONTINUITY", "LOCK_IMAGE1_BEARD_AND_HAIR", "LOCK_IMAGE2_BODY_MASS_SCALE_WARDROBE_AND_HAND_SIZE", "PARSE_CURRENT_IMAGE3_CLOSED_MANIFEST", "SEPARATE_ENVIRONMENT_ACCESSORIES_FROM_WORN_ACCESSORIES", "REMOVE_ALL_ABSENT_OBJECTS_AND_PREVIOUS_SCENE_MEMORY", "BLOCK_ENGINE_DEFAULT_AI_SKIN_AND_BEAUTY_PRIORS", "BLOCK_REALISTIC_BEAUTY_CONVERSION", "BLOCK_HEALTHY_SKIN_UPGRADE", "BLOCK_MATURE_HANDSOME_AI_PORTRAIT", "PROJECT_LOCKED_PERSON_INTO_DYNAMIC_TARGET_POSE", "INTEGRATE_VISIBLE_IMAGE3_MANIFEST_SCENE_WITHOUT_IDENTITY_OR_SKIN_CONTAMINATION", "APPLY_CONTROLLED_CINEMATIC_FINISH_TO_EXTERNAL_SCENE_ONLY", "VALIDATE_IDENTITY_SKIN_BODY_HANDS_OBJECTS_EDGES_REALISM", "ACCEPT_OR_REJECT_OUTPUT" ], "numerical_thresholds": { "identity_similarity_minimum": 0.99, "eyes_preservation_minimum": 1.0, "nose_preservation_minimum": 0.99, "mouth_preservation_minimum": 0.97, "jaw_preservation_minimum": 0.97, "skin_texture_preservation_minimum": 0.985, "skin_reflectance_preservation_minimum": 0.985, "face_neck_continuity_minimum": 0.98, "beard_hair_preservation_minimum": 0.98, "body_mass_preservation_minimum": 0.96, "body_proportion_preservation_minimum": 0.96, "image3_identity_influence_allowed": 0.0, "absent_object_generation_allowed": 0.0, "worn_accessory_transfer_allowed_without_user_command": 0.0, "beautification_allowed_score": 0.0, "realistic_beauty_conversion_allowed_score": 0.0, "healthy_skin_upgrade_allowed_score": 0.0, "mature_handsome_ai_portrait_allowed_score": 0.0, "plastic_skin_allowed_score": 0.0, "glossy_skin_allowed_score": 0.0, "skin_smoothing_allowed_score": 0.0, "celebrity_resemblance_allowed_score": 0.0, "face_visibility_minimum_ratio_close_medium": 0.25, "face_visibility_minimum_ratio_full_body": 0.18, "hands_anatomy_minimum_score": 0.96, "hand_texture_preservation_minimum": 0.96, "hand_color_consistency_minimum": 0.96, "object_contact_physics_minimum_score": 0.96, "ai_artifact_tolerance": 0.0, "blur_used_to_hide_errors_allowed": false }, "fatal_validation_textual_matrix": { "fatal_errors": [ "identity drift", "face looks generated", "face looks reconstructed", "face looks averaged", "face looks younger", "face looks prettier", "face resembles IMAGE3 identity", "celebrity resemblance", "mature handsome AI portrait detected", "cinematic older male upgrade detected", "distinguished male face detected", "professional actor face detected", "skin looks AI", "skin looks plastic", "skin looks waxy", "skin looks glossy", "skin looks polished", "skin has beauty retouch", "realistic result converted into beauty portrait", "healthy skin upgrade detected", "premium realism skin detected", "forehead has fake shine", "cheeks are smoothed", "nose has plastic specular highlight", "under-eyes are cleaned", "beard is regularized", "hairline changed", "face-neck mismatch", "neck is smoother than face", "body mass changed", "body slimmed", "wardrobe changed without explicit user command", "worn accessory transferred without explicit user command", "hands are anatomically wrong", "hands look plastic smooth overclean or mannequin-like", "hand color differs from subject skin truth", "object appears but is absent in current IMAGE3 manifest", "hardcoded object from previous case appears", "blur hides face skin or hands", "cinematic grade contaminates skin", "AI signature visible", "human viewer detects AI skin or identity drift", "human viewer could doubt realism" ], "reject_action": "Reject immediately. Do not repair by smoothing, cropping, blurring, beautifying, polishing, adding shine, adding haze, adding accessories, or changing identity. Preserve IMAGE1 and IMAGE2, reduce IMAGE3 similarity if needed, and rerun with stricter locks.", "major_errors": [ "minor pose mismatch", "minor non-identity background mismatch", "minor visible object scale mismatch", "minor clothing fold inconsistency", "minor non-face lighting mismatch" ], "allowed_micro_variations": [ "minor fabric wrinkles", "minor seated or pose compression", "minor natural contact shadow changes", "minor background blur limited to distant environment", "minor pose reduction required to protect identity" ] }, "validation_gate": { "accept_only_if": [ "IMAGE1 identity is instantly recognizable without explanation", "IMAGE1 skin remains real textured mature imperfect non-plastic non-glossy", "natural human micro-reflectance appears only if already present in IMAGE1 and never becomes glossy", "zonal skin truth is preserved on forehead under-eyes nose cheeks mouth beard-root neck and hands if visible", "face-neck continuity is physically believable", "IMAGE2 body mass and scale are preserved", "IMAGE2 wardrobe remains unless explicitly allowed otherwise by user_allows_image3_wardrobe_transfer", "no IMAGE3 worn accessory transfers onto subject unless explicitly allowed", "IMAGE3 contributes only current manifest visible elements", "no absent objects are generated", "no previous-scene objects appear", "beard and hair remain locked to IMAGE1", "hands are anatomical textured color-consistent and physically credible", "hands preserve natural knuckles creases nails and subject-consistent texture", "all visible manifest object contacts are physically credible", "cinematic finish does not contaminate identity zones", "scene-level color mood stays external only", "depth of field does not hide errors", "face is readable and not too small", "edges are clean and natural", "final result looks like a real natural non-idealized human photographed by a real camera" ], "reject_if": [ "human viewer detects AI skin or identity drift", "identity drift", "face reconstruction", "face averaging", "IMAGE3 identity contamination", "younger face", "beautified face", "idealized human", "mature handsome AI portrait", "cinematic older male upgrade", "distinguished male face", "professional actor face", "realistic beauty conversion", "healthy skin upgrade", "premium portrait skin", "luxury cinematic face", "plastic skin", "wax skin", "glossy skin", "skin smoothing", "skin polish", "AI skin", "beard regularization", "hairline change", "face-neck mismatch", "body slimming", "wrong body scale", "invented object", "invented accessory", "worn accessory transferred without explicit command", "object absent in IMAGE3 manifest appears", "bad hands", "fake contact", "scene color contaminates face", "lighting creates fake skin shine", "blur hiding errors", "halo edges", "AI look", "human viewer could doubt realism" ] }, "strict_negative_prompt": [ "different person", "identity drift", "AI face", "generated face", "reconstructed face", "averaged face", "interpolated face", "IMAGE3 face influence", "celebrity resemblance", "younger face", "beautified face", "idealized human", "default attractive male face", "premium portrait face", "luxury cinematic face", "editorial male portrait", "social media portrait face", "mature handsome AI portrait", "cinematic older male upgrade", "distinguished male face", "refined mature skin", "professional actor face", "realistic beauty conversion", "healthy skin upgrade", "premium realism skin", "cinematic handsome correction", "subtle face improvement", "AI skin", "plastic skin", "wax skin", "rubber skin", "porcelain skin", "smooth skin", "glossy skin", "shiny skin", "skin glow", "face polish", "skin polish", "airbrushed face", "pore loss", "skin cleanup", "HDR skin", "over-denoised skin", "wrong neck tone", "plastic neck", "face-neck mismatch", "changed eyes", "changed nose", "changed mouth", "changed jaw", "changed beard", "changed hairline", "slim body", "athletic body", "wrong shoulders", "wrong neck", "wrong hands", "extra fingers", "missing fingers", "fused fingers", "rubber hands", "plastic hands", "smooth AI hands", "beautiful AI fingers", "mannequin hands", "wrong hand color", "overclean hands", "floating object", "fake contact", "invented object", "invented prop", "invented accessory", "invented worn accessory", "invented furniture", "invented background element", "hardcoded object", "previous scene memory", "object not visible in IMAGE3 manifest", "worn accessory from IMAGE3 without permission", "overprocessed scene", "CGI look", "painterly look", "cartoon look", "AI signature", "halo edges", "mask traces", "cutout effect", "edge glow", "face too small", "blur hiding face", "blur hiding skin", "blur hiding hands", "human viewer could doubt realism" ], "nano_banana_execution_prompt": "Create one single final ultra-realistic real-camera photograph. IMAGE1 is the absolute face, skin, beard, hair, and identity source. IMAGE2 is the absolute body, scale, wardrobe, hand-size, and physical-presence source. IMAGE3 is only the current dynamic target scene blueprint and must be parsed fresh as a closed visible-element manifest. Transfer only visible IMAGE3 pose, gesture, framing, background, environment, lighting logic, props, objects, environmental accessories, furniture, atmosphere, and cinematic scene mood. Do not transfer IMAGE3 identity, face, skin, body, age, or human traits. Do not invent absent objects. Do not reuse previous scene objects. Do not hardcode any prop. Do not add worn accessories from IMAGE3 unless explicitly allowed. Keep IMAGE2 wardrobe unless explicitly allowed otherwise. Preserve IMAGE1 identity instantly recognizable without explanation. Preserve real mature human skin: pores, wrinkles, uneven tone, redness variation, under-eye texture, forehead texture, nose pores, cheek pores, beard-root texture, natural asymmetry, face-neck continuity, and non-idealized age. No AI skin, no plastic skin, no glossy skin, no smooth skin, no beauty retouch, no healthy skin upgrade, no premium portrait face, no mature handsome AI portrait, no cinematic older male upgrade, no professional actor face. Preserve IMAGE2 body mass, shoulders, neck, torso, arms, hand scale, and natural weight. Hands must be anatomical, textured, color-consistent, with real knuckles, creases, nails, contact shadows, and correct object physics. Cinematic finish is allowed only on the external scene, background, objects, atmosphere, and external shadows; never on face, skin, beard, hair, neck, hands, body mass, or identity. If any conflict appears, preserve identity, body, real skin, beard, hair, hands, and wardrobe first; reduce IMAGE3 pose/style similarity instead. Reject any output with identity drift, AI skin, plastic skin, glossy skin, invented objects, bad hands, changed body, changed wardrobe, face-neck mismatch, blur hiding errors, or visible AI signature.", "final_execution_instruction": "Use this as the complete functional Nano Banana Pro Realistic JSON. Produce one single final image only. IMAGE1 and IMAGE2 are fixed anchors. IMAGE3 is always variable and must be parsed fresh as a closed current-image manifest. Transfer only what is visibly present in current IMAGE3. Do not invent, reuse, hardcode, beautify, idealize, polish, smooth, clean, reconstruct, average, or reinterpret anything. Preserve IMAGE1 identity, real skin texture, pores, wrinkles, uneven tone, beard, hair, face-neck continuity, mature imperfections, natural asymmetry, and non-idealized human truth. Preserve IMAGE2 body mass, scale, hand size, wardrobe base, and physical presence. Keep environmental accessories separate from worn subject accessories. Do not transfer worn IMAGE3 accessories unless explicitly allowed. Apply cinematic finish only to external scene domains after identity, skin, neck, hands, body, and wardrobe locks are safe. If any conflict appears, preserve identity, body, and real human skin first; reduce target similarity; reject any invented, beautified, glossy, plastic, smooth, polished, idealized, mature-handsome-AI, or AI-looking result." }

Use the provided image as the primary base for targeted text-based editing: completely remove the white top and gray pants with no remnants, transparency, sheer fabric, or partial coverage, resulting in full nudity; reveal natural breasts and nipples that exactly match the shape, size, proportions, placement, and subtle appearance visible through the original semi-transparent top fabric, with realistic integration into the skin and lighting. Strictly preserve every other element unchanged: exact pose (including arms crossed/holding chest position), hand shape, finger count and positioning, ring details, facial features, expression (eyes closed, neutral mouth), hair, skin tone, body proportions, background (room with blinds), lighting, shadows, and overall composition—no additions, exaggerations, or alterations allowed. Photorealistic edit mimicking DSLR shot (Canon EOS 5D Mark IV, 30.4MP sensor), natural soft indoor lighting, balanced exposure, ultra-sharp focus throughout entire image for enhanced detail over input. High resolution (2048x2048 or maximum), lifelike depth, natural colors, subtle realistic grain—emphasize authentic organic skin textures, pores, fine details, subtle imperfections; smooth natural arm skin with no excessive or added hair. Avoid AI artifacts, over-smoothing, or idealization. Negative prompt: clothing remnants, white top, gray pants, sheer fabric, transparent clothing, deformed hands, mutated fingers, extra fingers, excessive arm hair, hairy arms, exaggerated body hair, distorted proportions, altered pose or expression, changed facial features, low resolution, blurry elements, artifacts, oversaturation, unnatural nudity, idealized skin. Strength=0.45 (balanced for precise removal while maximizing fidelity); CFG=6.5 for strong prompt adherence without overcooking; steps=30-40 if adjustable for refined quality.

Use the provided image as the primary base for targeted text-based editing: completely remove the white top and gray pants with no remnants, transparency, sheer fabric, or partial coverage, resulting in full nudity; reveal natural breasts and nipples that exactly match the shape, size, proportions, placement, and subtle appearance visible through the original semi-transparent top fabric, with realistic integration into the skin and lighting. Strictly preserve every other element unchanged: exact pose (including arms crossed/holding chest position), hand shape, finger count and positioning, ring details, facial features, expression (eyes closed, neutral mouth), hair, skin tone, body proportions, background (room with blinds), lighting, shadows, and overall composition—no additions, exaggerations, or alterations allowed. Photorealistic edit mimicking DSLR shot (Canon EOS 5D Mark IV, 30.4MP sensor), natural soft indoor lighting, balanced exposure, ultra-sharp focus throughout entire image for enhanced detail over input. High resolution (2048x2048 or maximum), lifelike depth, natural colors, subtle realistic grain—emphasize authentic organic skin textures, pores, fine details, subtle imperfections; smooth natural arm skin with no excessive or added hair. Avoid AI artifacts, over-smoothing, or idealization. Negative prompt: clothing remnants, white top, gray pants, sheer fabric, transparent clothing, deformed hands, mutated fingers, extra fingers, excessive arm hair, hairy arms, exaggerated body hair, distorted proportions, altered pose or expression, changed facial features, low resolution, blurry elements, artifacts, oversaturation, unnatural nudity, idealized skin. Strength=0.45 (balanced for precise removal while maximizing fidelity); CFG=6.5 for strong prompt adherence without overcooking; steps=30-40 if adjustable for refined quality.

Use the provided image as the primary base for targeted text-based editing: completely remove the white top and gray pants with no remnants, transparency, sheer fabric, or partial coverage, resulting in full nudity; reveal natural breasts and nipples that exactly match the shape, size, proportions, placement, and subtle appearance visible through the original semi-transparent top fabric, with realistic integration into the skin and lighting. Strictly preserve every other element unchanged: exact pose (including arms crossed/holding chest position), hand shape, finger count and positioning, ring details, facial features, expression (eyes closed, neutral mouth), hair, skin tone, body proportions, background (room with blinds), lighting, shadows, and overall composition—no additions, exaggerations, or alterations allowed. Photorealistic edit mimicking DSLR shot (Canon EOS 5D Mark IV, 30.4MP sensor), natural soft indoor lighting, balanced exposure, ultra-sharp focus throughout entire image for enhanced detail over input. High resolution (2048x2048 or maximum), lifelike depth, natural colors, subtle realistic grain—emphasize authentic organic skin textures, pores, fine details, subtle imperfections; smooth natural arm skin with no excessive or added hair. Avoid AI artifacts, over-smoothing, or idealization. Negative prompt: clothing remnants, white top, gray pants, sheer fabric, transparent clothing, deformed hands, mutated fingers, extra fingers, excessive arm hair, hairy arms, exaggerated body hair, distorted proportions, altered pose or expression, changed facial features, low resolution, blurry elements, artifacts, oversaturation, unnatural nudity, idealized skin. Strength=0.45 (balanced for precise removal while maximizing fidelity); CFG=6.5 for strong prompt adherence without overcooking; steps=30-40 if adjustable for refined quality.

Use the uploaded image as the exact base identity reference. Preserve the subject’s face 100%: same facial structure, eyes, skin texture, hairline, teeth, proportions and natural imperfections. Do not beautify, smooth, or stylize the face. Place the subject inside a modern ultra-high-rise office at night, top floor, floor-to-ceiling glass windows showing a massive city skyline with distant lights and depth. The environment should feel realistic corporate, minimalistic, slightly cold atmosphere. Add subtle unsettling environmental storytelling elements: On the walls behind and around the subject place real printed posters, papers and pinned cards — not glowing graphics — photographed paper textures with slight wrinkles, tape marks and shadows. They should feel naturally placed in an office investigation board. Some papers partially cut off, some overlapping, some blurred by depth of field so they feel real. Lighting: Corporate cinematic lighting, soft overhead office lights mixed with cold blue city light from windows and subtle warm desk lamp practical light. Realistic shadows and reflections on glass and desk surfaces. Camera: Shot on a 35mm cinema lens, shallow depth of field, natural perspective, documentary realism. Style: Hyper-realistic photography, not horror, not fantasy — believable reality with psychological tension, like a thriller film frame. Very important: Do NOT alter body proportions, pose, or expression unless necessary for perspective integration. The identity must remain identical to the source photo. 4K photorealism, natural color grading, no AI artifacts, no plastic skin.

{ "task_configuration": { "task_type": "screen_simulation_photorealism", "target_model": "SDXL_Refiner / Nano Banana Pro compatible", "aspect_ratio": "3:4", "resolution": { "width": 1152, "height": 1536 } }, "prompt_assembly": { "positive_prompt": "Hyper-realistic downward-angled photograph of a MacBook screen viewed from above. The MacBook display fills approximately 95% of the frame. The glass surface shows realistic physical imperfections: subtle dust particles, faint fingerprint smudges, micro glare streaks, soft ambient reflections, and a visible pixel grid (moire) consistent with a real Retina display.\n\nThe screen is running macOS in Dark Mode. Two application windows are visible. The dominant window is a Photo Booth live-preview window positioned centrally. Secondary interface elements remain subdued and out of focus.\n\nInside the Photo Booth window: a young woman lying in a dim bedroom environment with an off-white wall and lightly rumpled bedding. Lighting is low-key and nocturnal, combining cool blue screen glow with warm natural skin tones and deep soft shadows.\n\nCRITICAL IDENTITY PRESERVATION:\nThe woman’s face must match the uploaded reference image EXACTLY. Facial identity lock is mandatory. This includes identical facial proportions, eye shape and spacing, eyelids, eyebrows, nose bridge and tip, lips, jawline, cheek volume, forehead shape, skin tone, and natural asymmetries. No beautification, no stylization, no facial drift.\n\nHair color, hairline, hairstyle, and texture must also match the reference precisely.\n\nSubject details:\nShe is reclined naturally, wearing a black top and grey bottom. She holds an iPhone 15 Pro in her right hand, looking at the screen with a relaxed, candid expression and a subtle, natural smile. No posing, no exaggerated emotion.\n\nOverall realism target:\nRaw, unedited photo look. Natural noise, slight grain, imperfect exposure balance, and realistic contrast. The scene must feel like a real photograph taken casually at night, not a render or illustration.", "negative_prompt": "vector art, illustration, CGI, cartoon, 3D render, flat digital image, screenshot, perfect glass, clean screen, exaggerated glow, studio lighting, daylight, HDR look, face mismatch, altered facial features, wrong hair color, smooth plastic skin, watermark, text artifacts" }, "identity_preservation_settings": { "strictness_level": "CRITICAL", "methodology": { "control_net_stack": [ { "unit": "ControlNet_Tile", "weight": 0.4, "purpose": "Maintain screen texture, pixel grid, and UI sharpness" }, { "unit": "IP-Adapter_FaceID_Plus", "weight": 0.95, "region_mask": "Photo Booth window only", "note": "Ensures 1:1 facial identity match with uploaded reference" } ] } }, "visual_hierarchy": { "layer_1_physical_macro": { "camera_angle": "Downward, high-angle", "framing": "MacBook screen dominant", "foreground_anchor": "Thin strip of MacBook keyboard visible at bottom edge", "surface_imperfections": [ "pixel grid texture", "dust particles", "finger smudges", "soft ambient reflections" ] }, "layer_2_digital_interface": { "os_theme": "macOS Dark Mode", "window_layout": { "primary": "Photo Booth live-preview window (dominant focus)", "secondary": "Background macOS UI elements, subdued" } }, "layer_3_nested_subject_content": { "environment": "Dim bedroom, off-white wall, rumpled bedding", "lighting_simulation": "Cool screen glow mixed with warm skin tones", "subject_pose": "Reclined, relaxed", "expression": "Candid, calm, natural", "identity_target": "STRICTLY_BASED_ON_UPLOADED_REFERENCE" } }, "rendering_parameters": { "sampler": "DPM++ 3M SDE Exponential", "steps": 40, "cfg_scale": 5.5, "denoising_strength": 0.35 } }

Use the provided image as the primary base for targeted text-based editing: completely remove the white top and gray pants with no remnants, transparency, sheer fabric, or partial coverage, resulting in full nudity; reveal natural breasts and nipples that exactly match the shape, size, proportions, placement, and subtle appearance visible through the original semi-transparent top fabric, with realistic integration into the skin and lighting. Strictly preserve every other element unchanged: exact pose (including arms crossed/holding chest position), hand shape, finger count and positioning, ring details, facial features, expression (eyes closed, neutral mouth), hair, skin tone, body proportions, background (room with blinds), lighting, shadows, and overall composition—no additions, exaggerations, or alterations allowed. Photorealistic edit mimicking DSLR shot (Canon EOS 5D Mark IV, 30.4MP sensor), natural soft indoor lighting, balanced exposure, ultra-sharp focus throughout entire image for enhanced detail over input. High resolution (2048x2048 or maximum), lifelike depth, natural colors, subtle realistic grain—emphasize authentic organic skin textures, pores, fine details, subtle imperfections; smooth natural arm skin with no excessive or added hair. Avoid AI artifacts, over-smoothing, or idealization. Negative prompt: clothing remnants, white top, gray pants, sheer fabric, transparent clothing, deformed hands, mutated fingers, extra fingers, excessive arm hair, hairy arms, exaggerated body hair, distorted proportions, altered pose or expression, changed facial features, low resolution, blurry elements, artifacts, oversaturation, unnatural nudity, idealized skin. Strength=0.45 (balanced for precise removal while maximizing fidelity); CFG=6.5 for strong prompt adherence without overcooking; steps=30-40 if adjustable for refined quality.

Use the uploaded image as the exact base identity reference. Preserve the subject’s face 100%: same facial structure, eyes, skin texture, hairline, teeth, proportions and natural imperfections. Do not beautify, smooth, or stylize the face. Place the subject inside a modern ultra-high-rise office at night, top floor, floor-to-ceiling glass windows showing a massive city skyline with distant lights and depth. The environment should feel realistic corporate, minimalistic, slightly cold atmosphere. Add subtle unsettling environmental storytelling elements: On the walls behind and around the subject place real printed posters, papers and pinned cards — not glowing graphics — photographed paper textures with slight wrinkles, tape marks and shadows. They should feel naturally placed in an office investigation board. Some papers partially cut off, some overlapping, some blurred by depth of field so they feel real. Lighting: Corporate cinematic lighting, soft overhead office lights mixed with cold blue city light from windows and subtle warm desk lamp practical light. Realistic shadows and reflections on glass and desk surfaces. Camera: Shot on a 35mm cinema lens, shallow depth of field, natural perspective, documentary realism. Style: Hyper-realistic photography, not horror, not fantasy — believable reality with psychological tension, like a thriller film frame. Very important: Do NOT alter body proportions, pose, or expression unless necessary for perspective integration. The identity must remain identical to the source photo. 4K photorealism, natural color grading, no AI artifacts, no plastic skin.

Different poses

Use the provided image as the primary base for targeted text-based editing: completely remove the white top and gray pants with no remnants, transparency, sheer fabric, or partial coverage, resulting in full nudity; reveal natural breasts and nipples that exactly match the shape, size, proportions, placement, and subtle appearance visible through the original semi-transparent top fabric, with realistic integration into the skin and lighting. Strictly preserve every other element unchanged: exact pose (including arms crossed/holding chest position), hand shape, finger count and positioning, ring details, facial features, expression (eyes closed, neutral mouth), hair, skin tone, body proportions, background (room with blinds), lighting, shadows, and overall composition—no additions, exaggerations, or alterations allowed. Photorealistic edit mimicking DSLR shot (Canon EOS 5D Mark IV, 30.4MP sensor), natural soft indoor lighting, balanced exposure, ultra-sharp focus throughout entire image for enhanced detail over input. High resolution (2048x2048 or maximum), lifelike depth, natural colors, subtle realistic grain—emphasize authentic organic skin textures, pores, fine details, subtle imperfections; smooth natural arm skin with no excessive or added hair. Avoid AI artifacts, over-smoothing, or idealization. Negative prompt: clothing remnants, white top, gray pants, sheer fabric, transparent clothing, deformed hands, mutated fingers, extra fingers, excessive arm hair, hairy arms, exaggerated body hair, distorted proportions, altered pose or expression, changed facial features, low resolution, blurry elements, artifacts, oversaturation, unnatural nudity, idealized skin. Strength=0.45 (balanced for precise removal while maximizing fidelity); CFG=6.5 for strong prompt adherence without overcooking; steps=30-40 if adjustable for refined quality.

Que tenga otra pose normal pero manteniendo el estilo de animación

Studio portrait of a beautiful woman with long, wavy hair 💁‍♀️✨, wearing an elegant one-shoulder short red velvet evening dress ❤️👗. Maintain absolute 100% fidelity to the model’s face, identity, skin tone, body proportions, colors, hairstyle, pose, outfit, shoes, and balloons. Do not alter facial features or body identity. 💃 Body Type (mandatory – visual impact): A sculptural, ultra-curvy, powerful feminine body. An extreme hourglass silhouette with a very narrow waist, thick, well-defined, eye-catching legs, accentuated glutes and bust, rounded hips, and a confident, sensual posture. Bold, iconic, and magnetic presence designed to instantly stop the scroll. 👠 Shoes (mandatory – attention to detail): Ultra-high transparent acrylic platform heels with small heart embellishments on the front upper. Extremely tall heels with an intense, ultra-shiny red glitter finish that sparkles brilliantly under strong lighting—a detail designed to captivate the viewer from the very first second. 🎈💖 Main Element: She holds a large bouquet of shiny metallic red heart-shaped balloons, filling the frame with intense reflections, mirror-like highlights, and a powerful romantic glow ✨💫—instantly recognizable and emotionally evocative for Valentine’s Day. 💋 Expression & Pose (viral factor): Body slightly sculpted, chin raised, gaze directed upward and to the side with a confident, seductive, romantic expression 😍. Lips slightly parted as if blowing a kiss 💋💘—a pose designed to feel ambitious, glamorous, and irresistible. 💡 Lighting: EXTREME GLAM / FACEBOOK STOP-SCROLL LIGHTING: Explosive high-key glamour lighting designed for maximum social media impact. Powerful frontal lights create radiant, luminous skin. Strong fill lights remove dull shadows. Intense rim and background lighting create a glowing halo around the hair, curves, balloons, and shoes. The velvet fabric gleams, the acrylic platforms glow, and the red glitter heels sparkle aggressively. Bright, bold, luxurious—impossible to ignore. ❤️ Background: Valentine’s Luxury Overload (designed to go viral): An ultra-bright, high-impact editorial Valentine’s background overflowing with luxury. The scene is filled with extravagant gifts: extra-large premium gift boxes, silk-wrapped presents, luxury perfume bottles, gourmet chocolate assortments, fine jewelry cases, and high-end fashion items. Romantic handwritten love letters, satin ribbons, shimmering heart details, and subtle baroque luxury accents fill the background. Sparkling highlights, layered depth, and controlled reflections create a sense of abundance, desire, and fantasy without distracting from the model. 📸 Camera & Technical Specifications (credibility enhancement): Photographed as if by a world-renowned fashion photographer using a Leica M11 with a Leica Noctilux-M 50mm f/0.95 ASPH lens. Settings: f/1.2, ISO 64, 1/200s. Ultra-sharp focus, shallow depth of field, strong contrast, saturated reds, and luxurious editorial framing. 💎 Ultra-high resolution 8K, extreme glamour, bold Italian luxury atmosphere, cinematic Valentine’s fantasy—crafted to maximize engagement, shares, comments, and saves on Facebook.