Versions: Pony V1 & XL V1
Cheeky is a stylized LoRA designed to inject playful confidence, expressive posing, spunk, and bold charm into your anime-style generations. Built around a distinct attitude, this LoRA emphasizes over-the-shoulder glances, smirks, and that unmistakable βyou caught me on purposeβ energy.
Whether used subtly or cranked to full power, Cheeky turns your characters into scene-stealers β teasing, charismatic, and dripping with visual appeal.
There are currently two versions available:
Cheeky V1 (Pony) β Tailored for pony/anthro models with great expressiveness and dynamic posing.
CheekyXL V1 (SDXL) β Designed for SDXL anime models but surprisingly effective with realism and crossover models too.
Clean, anime-style linework
Soft, glossy shading with subtle highlights
Teasing facial expressions (smirks, flirty glances)
Suggestive, confident poses from behind
Balanced fanservice without overdoing it
Great for semi-pinup, character portraits, or expressive close-ups
Trigger Words:
cheekystyle
(Pony V1)
cheekyxl
(XL V1)
LoRA Strength:
1.0 β 1.2
: Balanced aesthetic
1.4 β 1.6
: Greater stylization, bolder lighting and gloss
1.7 β 1.9
: Max stylization, almost full 2D anime rendering
Note: The XL versionβs sweet spot is around 1.0, but can go up to 1.9 depending on the model and taste.
Sampler:
Euler
Euler a
DPM++ 2M
(Any anime-compatible sampler works well)
Scheduler:
Karras
Others: Not yet tested β feel free to experiment and share results!
Steps:
15+ recommended
Higher step counts yield better detail and lighting definition
CFG:
3.0 β 8.0
recommended for soft, clean outputs
Other values may also work β feel free to experiment
Adetailer:
Pony version: Strongly recommended for clean eye renders
XL version: Model-dependent β many SDXL bases already produce great eyes out of the box; use Adetailer only if needed
π Poses & Expressions:
from behind
looking back
looking at viewer
ass (as a pose/composition angle)
flirty pose
smirk
teasing expression
over the shoulder
π¨ Style & Lighting:
cleavage, glossy skin, dynamic angle, illustration, studio light, soft bokeh, sunset lighting, golden hour, cinematic light
While Cheeky focuses on flirtation and expression, it can produce NSFW content depending on your base model and prompt. When paired with permissive or explicit-capable models, it may emphasize anatomy and body-centric poses more heavily. Use responsibly based on your audience and platform.
Anime-style female character renders
Expressive pin-up portraits
Mood-rich teaser scenes
Flirty character sheets or commissions
Realism models with a cheeky visual edge
Pony/anthro models needing a spark of personality
Cheeky V1 (Pony): Best with pony/anthro-focused base models
CheekyXL V1: Best with SDXL anime-style bases
Also works well with realism and crossover SDXL models
On realism bases, you'll get more grounded anatomy with a flirtatious tone
On anime or pony bases, expect fully stylized output with a polished 2D finish
(Model-dependent β adjust based on your base model's style)
Prompt 1 (Anime):anime, 1girl, 25-year-old woman, from behind, looking back at viewer, wearing a lilac sundress, smirk, flirty pose, cleavage, glossy skin, dynamic angle, golden hour light, soft bokeh, illustration, in the style of cheekyxl <lora:CheekyXL-V1_epoch_10:1.0>
Prompt 2 (Anime):anime, 1girl, in a tight red crop top and shorts, ass, looking back, smirking, confident pose, messy short hair, rim light, studio light, cheekyxl aesthetic <lora:CheekyXL-V1_epoch_10:1.3>
Prompt 3 (Anime Pinup):anime, 1girl, bunny suit, kneeling from behind, looking at viewer over shoulder, teasing smile, cleavage, glossy thighs, dramatic spotlight, soft vignette, cheekyxl <lora:CheekyXL-V1_epoch_10:1.6>
Prompt 4 (Pony):pony, mare, from behind, looking back, smirk, flirty pose, flowing mane, cutie mark visible, sunset light, dynamic angle, cinematic bokeh, cheekystyle <lora:Cheeky_V1_epoch_10:1.2>
Prompt 5 (Pony Portrait):pony, unicorn, cheeky expression, over the shoulder glance, twinkling eyes, glossy coat, studio lighting, in the style of cheekystyle <lora:Cheeky_V1_epoch_10:1.0>
This LoRA is strictly for personal, non-commercial use only.
You are not allowed to repost, sell, merge, or otherwise profit from this model in any form.
It is allowed on CivitAI for personal use only. Unauthorized redistribution or monetization is prohibited.
βοΈ CheekyXL V1 β Settings Summary
π Trigger Word: cheekyxl
π Recommended Strength:
Sweet spot: 1.0
Range: 1.0 β 1.9
(more stylized with higher values)
π§ͺ Samplers:
Euler
Euler a
DPM++ 2M
(Any SDXL-compatible anime sampler should work well)
π Scheduler:
Karras
(Others untested β feel free to experiment)
π Steps:
15+ recommended
More steps = better lighting and detail refinement
ποΈ CFG Scale:
3.0 β 8.0
for soft, clean, aesthetic outputs
Tune based on prompt/model and desired contrast or boldness
ποΈ Adetailer:
Model-dependent
Most SDXL bases generate excellent eyes natively
Use only if your base model benefits from it
Go ahead and upload yours!
Your query returned no results β please try removing some filters or trying a different term.