GPT Image Prompt
Same real person from the active LoRA identity reference, preserving her exact recognizable facial geometry and full body identity as the highest priority: same face shape, eye spacing, nose/lip/jaw/cheek structure, skull proportions, long dark loose hair identity, light skin tone, shoulder-to-head relationship, shoulder width, torso length, torso-to-leg balance, natural waist thickness, hip width, hip-to-thigh transition, thigh volume, leg length ratio, grounded adult height impression, and natural body mass distribution. Do not redesign her into a slimmer, taller, curvier, more model-like, more influencer-like, or more cosmetically polished version. Preserve the LoRA reference body geometry exactly even under the outfit, pose, crop, lighting, and smartphone perspective.
She is standing indoors near bright sheer white curtains in an ordinary real room, leaning slightly forward with her torso subtly angled, head tilted down to the side, arms relaxed forward, and both hands gathered low near the front of her skirt. Her posture should feel naturally caught in a quiet in-between moment, not posed like fashion photography. Her expression stays subtle and alive: a soft camera-aware glance, relaxed eyelids, slightly parted lips, mild asymmetry in the mouth and eyes, and a calm human presence as if the phone photo caught her between moments rather than holding a frozen pose. Do not copy any blank, stiff, dead-eyed, or mannequin-like expression from the reference if the reference is neutral; keep the LoRA facial structure while applying this soft living expression.
She wears an oversized white button-up shirt tied or lifted casually at the front, with long loose sleeves and realistic cotton wrinkles, a narrow black necktie hanging down naturally, and a dark pleated mini skirt with believable fabric folds, weight, and slight unevenness. The clothing must sit on and reveal the same LoRA body proportions underneath, without compressing the waist, narrowing the hips, reducing thigh volume, elongating the legs, or reshaping her silhouette into a generic attractive body type. Hands gathered low near the skirt must have correct finger count, natural finger overlap, believable wrist flow, realistic pressure where fingers meet fabric, and no fused or duplicated fingers.
The pose, expression, outfit, gesture, indoor curtain setting, camera behavior, lighting, and realism direction all remain subordinate to the active LoRA identity reference. The final subject must still be unmistakably the exact same LoRA person in both face and body, not a similar-looking substitute, not a beautified replacement, not a generic attractive version, and not the same face placed on a different body. Preserve the source-prompt human moment while protecting LoRA identity structure.
Casual iPhone-style vertical photo at close indoor portrait distance, imperfect handheld framing, slightly asymmetrical composition, mild lower-body crop, realistic smartphone wide-angle behavior without body-slimming distortion, and ordinary social-photo timing. Bright window daylight through sheer white curtains creates uneven highlights, partial clipping on the white shirt, soft shadow patches across shirt and skin, realistic indoor exposure compromise, mild HDR reconstruction, small-sensor depth behavior, and no cinematic lighting control. Allow realistic smartphone artifacts: mild compression softness, uneven sharpness in hair and fabric, localized skin softness, subtle edge reconstruction instability, slight exposure imbalance, natural flyaways, wrinkled cotton texture, imperfect fabric detail, and real-life visual clutter or asymmetry. The image should feel like a believable handheld iPhone capture, not a studio portrait, not a fashion editorial, not an influencer post, not a cinematic scene, and not a polished AI render.
Negative restrictions: no beauty-filter redesign, no influencer face, no seductive model gaze, no commercial smile, no porcelain skin, no waxy skin, no excessive facial symmetry, no body slimming, no waist compression, no hip narrowing, no thigh reduction, no leg elongation, no fashion-model proportions, no generic attractive-body substitution, no stiff mannequin expression, no blank stare, no dead eyes, no frozen mouth, no editorial posing, no luxury lifestyle styling, no cinematic color grading, no DSLR bokeh, no over-clean HDR, no globally sharp rendering, no plastic hands, no fused fingers, no duplicated fingers, no broken wrists, no floating body, no disconnected shadows, and no artificial perfection.
How to use: Copy and paste this prompt into ChatGPT, Gemini, Nano Banana, or your favorite AI image tool. If it mentions LoRA, reference image, uploaded photo, or identity reference, upload your own photo first. No reference photo? Replace the character with a clear age, gender, and nationality, such as “25-year-old Japanese woman” or “30-year-old American man”.