GPT Image Prompt
Use the uploaded portrait as the only active LoRA identity reference and preserve the exact same person’s structural identity: real facial geometry, face shape, eye spacing, eye shape, nose structure, lips, jawline, cheek structure, skull proportions, natural facial asymmetry, skin tone, hair identity, and recognizable physical presence. The uploaded portrait controls identity only; do not copy its original clothing, pose, background, lighting, emotional stiffness, or photographic style. Preserve the subject as the exact same person, not a beautified substitute, not a generic fashion model, and not a similar-looking replacement.
Create an ultra-realistic vertical 9:16 smartphone-captured surreal editorial beauty portrait where the subject’s face is partially visible behind a narrow, irregular vertical torn-paper opening, as if seen through layered ripped blue paper. The opening should reveal only part of the face, with the subject close behind the paper, creating a mysterious poetic beauty-portrait moment while keeping the identity clearly recognizable. The torn paper edges must look rough, natural, fibrous, uneven, tactile, and physically layered, with believable thickness, small paper fibers, tiny imperfections, and realistic edge shadows. The paper must feel like real material photographed up close, not a flat graphic design.
The subject has a calm, elegant, living expression with closed or softly lowered eyes, relaxed eyelids, subtle natural facial asymmetry, and a quiet refined emotional presence. The mouth remains soft and natural, with soft red-toned lips as the main color accent. Makeup is refined and minimal editorial beauty makeup, enhancing the real facial features without changing identity. Skin must remain smooth but realistic, with visible natural pores, subtle texture variation, realistic tonal transitions, and no plastic smoothing, no poreless beauty-filter effect, no waxy skin, no AI-polished face. Keep the face delicate and clean, but structurally faithful to the uploaded identity reference.
Use a cool blue monochrome palette across the torn paper layers and background, with the red-toned lips and one pale pink flower as restrained color accents. Lighting should feel like a premium surreal editorial studio setup but still physically believable: strong diagonal light and shadow crossing the blue paper surface, soft but high-contrast falloff, realistic shadow attachment inside the torn opening, and believable depth between foreground paper, face, and flower. Avoid fake HDR perfection; preserve natural exposure variation, subtle softness, slight reconstruction inconsistency, and realistic photographic texture.
Place a single delicate artistic flower near the lower side of the torn vertical opening. The flower has one pale pink blossom and a dark sculptural stem, minimal and poetic, not crowded, not decorative clutter. It should cast realistic small shadows onto the blue paper and feel physically placed near the torn edge, with believable scale, contact, and depth. No extra flowers.
Clothing must not copy the uploaded portrait. Use only a barely visible elegant minimal high-fashion garment, mostly hidden behind the torn paper opening, chosen to support the face, paper texture, and flower without distracting. Clothing should be understated, refined, and physically plausible, with no logo, text, or brand mark.
All expression, gesture, outfit, pose, camera, lighting, paper design, flower placement, editorial styling, and realism behavior remain subordinate to the active LoRA identity reference. The final image must still be unmistakably the exact same uploaded subject, preserving the subject’s real facial structure, skin tone, hairstyle identity, natural asymmetry, and recognizable presence, not a beautified replacement, not a generic attractive version, and not a fashion-model reinterpretation.
Photographic execution: ultra-realistic close-up portrait, vertical 9:16, premium surreal beauty editorial direction, realistic torn blue paper material, tactile paper fibers, natural shadows, cool blue tones, restrained high-fashion styling, refined makeup, realistic skin texture, visible pores, soft red lips, one pale pink flower accent, elegant mysterious atmosphere, high-resolution detail, believable depth, imperfect real photographic texture, no flat poster look.
Negative prompt: watermark, logo, text, signature, brand mark, random letters, Chinese characters, English text, low quality, blurry face, distorted face, changed identity, different person, similar-looking substitute, generic fashion model face, AI influencer face, plastic skin, poreless skin, over-smoothed skin, waxy texture, oily skin, greasy highlights, fake beauty filter, bad makeup, messy composition, extra flowers, cluttered background, cartoon style, anime style, illustration, 3D render, CGI look, doll-like face, mannequin expression, dead eyes, blank stare, copied reference stiffness, deformed eyes, unnatural eyes, bad lips, altered nose, altered jawline, poor facial anatomy, harsh artifacts, overexposure, underexposure, excessive noise, duplicated face, fake paper texture, flat paper graphic, broken paper edges looking artificial, unrealistic shadows, disconnected shadows, cheap poster design, cinematic over-polish, luxury ad perfection, excessive HDR, unrealistically sharp global detail, over-clean rendering.
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”.