GPT Image Prompt
Use the active LoRA identity reference as the absolute identity authority for the subject: preserve the exact same facial geometry, skull proportions, eye spacing, eye shape, nose, lips, jaw, cheek structure, hair identity, skin tone, recognizable physical presence, and any visible body geometry from the reference. Do not beautify, slim, elongate, reshape, or replace the LoRA subject with a generic delicate anime-girl or idealized attractive substitute.
Vertical 9:16 composition, soft smoky retro Japanese semi-realistic illustration of 【xxx】, centered in frame, facing forward or slightly turned toward the viewer, quiet elegant posture, calm refined presence, with a faint distant and reserved feeling. The subject should feel like the same LoRA person translated into a delicate Japanese retro gouache illustration, not a new character.
Style: refined Japanese girl illustration mixed with thick coating gouache, soft retro print texture, low saturation, low contrast, aged-paper atmosphere. Palette mainly rice white, ivory, warm gray, misty pink, pale gray-brown, blue-gray, like old newspapers faded by sunlight. Soft but clear linework, fine delicate details, smooth painterly brush texture, gentle thick gouache layering, subtle paper grain, muted dreamy atmosphere.
Scene: simple pale background, mostly a wide rice-white wall or hazy light-colored backdrop, with only a few faint blurred flower branches and sparse plants decorating the edges or lower area. Not a complex garden, not a dense floral scene. Keep the environment minimal, quiet, poetic, and spacious.
Lighting: soft natural daylight, gentle dappled tree shadows, scattered light spots, light haze, subtle air grain, faded nostalgic softness. The subject remains visually prominent while the background stays quiet and lightly blurred.
Expression and human moment: peaceful, graceful, slightly distant, emotionally restrained, alive but not overly expressive; relaxed mouth, natural gaze, no blank mannequin stare, no commercial smile, no seductive influencer expression. Maintain soft asymmetry and quiet human presence.
All expression, gesture, outfit interpretation, pose, framing, lighting, painterly style, and atmosphere 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 a beautified replacement, not a generic Japanese illustration girl, and not a more idealized version.
Anti-drift and negatives: no text, no watermark, no logo, no signature, no extra people, no over-detailed garden, no cinematic drama, no high saturation, no harsh contrast, no glossy digital anime finish, no plastic skin, no perfect symmetry, no AI-polished beauty, no fashion-editorial posing, no influencer gaze, no exaggerated body proportions, no waist slimming, no leg elongation, no hip or thigh reduction, no face reshaping, no dead eyes, no frozen neutral face, no copied reference stiffness.
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”.