GPT Image Prompt
Same real woman from the active LoRA identity reference images, preserving her exact recognizable facial geometry, fair-light skin tone, long dark brown hair with natural center-parted fall, same shoulder-to-head relationship, narrow shoulder width, torso length, torso-to-leg balance, waist thickness, hip width, hip-to-thigh transition, thigh volume, leg length ratio, knee-to-ankle scale, body mass distribution, and grounded adult physical presence. Do not reuse the previous generated woman. Do not shift her into a warmer medium-brown skin tone, tied-back hair, different face, thicker body, shorter body ratio, or generic AI courtyard subject. The final subject must remain unmistakably the exact same LoRA woman from the uploaded reference set in both face and full-body geometry.
She is standing naturally on outdoor garden stone steps in a light green matching summer set: long-sleeve tie-front crop top with a gathered front knot and matching high-waisted shorts. The outfit must follow her actual LoRA body silhouette without compressing her waist, widening or narrowing her hips, reducing or enlarging thigh volume, elongating her legs, shortening her torso, or creating a different body type. Delicate necklace and small earrings are visible. A small black handbag hangs from her relaxed left hand with believable grip pressure and natural finger spacing; her right hand rests lightly on the black metal railing with realistic palm contact, wrist angle, and pressure. Her feet and body weight feel physically grounded on the stone steps, with believable standing balance, realistic posture asymmetry, contact shadows, and natural weight transfer.
Her head is tilted back slightly into the bright sun, eyes closed naturally from the light, lips softly closed, expression calm and quietly pleased, like a casual phone photo caught her pausing for one second rather than posing stiffly. Preserve this living human moment on the LoRA subject while keeping her facial structure intact. Do not copy a blank studio-reference stare, dead-eyed neutrality, mannequin mouth, fashion-model gaze, commercial smile, seductive expression, or theatrical emotion. Expression should feel subtle, relaxed, observed, and real.
Keep the white stucco building, arched doorway, stone steps, potted greenery, black railing, and red flowering vines around her. Real leaf shadows fall unevenly across her face, neck, skin, arms, legs, and light green fabric. The environment feels physically measurable and ordinary: believable stair scale, realistic railing position, practical garden clutter, imperfect plant overlap, natural asymmetry, uneven background spacing, and authentic human-to-environment grounding. Lighting interacts consistently with the architecture, foliage, fabric, skin, handbag, jewelry, and shadows.
All expression, gesture, outfit, pose, camera behavior, lighting, environment, realism texture, smartphone processing, and aesthetic choices remain subordinate to the active LoRA identity reference. The final image must still be unmistakably the exact same LoRA subject, not a similar-looking substitute, not the previous failed generated subject, not a beautified replacement, not a different ethnicity or skin tone, not a different hair identity, not a slimmer or curvier reinterpretation, not a fashion-model version, and not a generic attractive version of the source prompt.
Casual iPhone-style vertical snapshot, close social-photo distance, slight handheld imperfection, mild casual tilt, imperfect crop balance, realistic wide-angle smartphone perspective that does not alter her body proportions. Bright midday sunlight with hard highlights, deeper foliage shadows, slightly blown highlight areas, imperfect smartphone HDR, uneven exposure compromise, localized sharpness variation, mild compression softness, subtle edge reconstruction artifacts, realistic skin texture, natural fabric wrinkles, believable jewelry and handbag detail. No cinematic lighting, no DSLR blur, no luxury editorial polish, no influencer-photo composition, no perfect symmetry, no airbrushed skin, no plastic texture, no over-clean AI rendering, no artificial depth isolation, no impossible architecture, no floating feet, no disconnected shadows, no fused fingers, no duplicated fingers, no broken wrists, no fake grip, no waist slimming, no hip reshaping, no thigh reshaping, no leg elongation, no body-type substitution, no skin-tone drift, no hair-identity drift, no face replacement.
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”.