
Loading…
Loading…

Creator · johnAGI168
View original source ↗九宫格人物站位图📷
prompt 👇
Using the uploaded reference photo as the single source of truth for the person, create a high-resolution cinematic fashion contact sheet: a 3×3 grid of nine vertical panels on a vertical 9:16 canvas, each panel also a vertical 9:16 frame, separated by black borders with softly rounded panel corners. The person in every panel must be exactly the person in the reference photo, reproduced with 100% fidelity: identical face shape and bone structure, eyes, eyebrows, nose, lips, skin tone and texture, freckles and moles; identical hairstyle, hair length, hairline and hair color; identical makeup; identical body build and proportions. The outfit, shoes, accessories and jewelry must match the reference exactly in style, color, material and fit. Do not beautify, slim, de-age or restyle anything. Take only the person from the reference, never its background.
Every panel is shot in the same completely clean, seamless light-gray infinity studio, where the floor curves into the wall with no visible horizon line, no corners, no props, no furniture and no architecture. The only things in the frame are the person and their soft natural shadow on the floor. Lighting is soft, even, diffused studio light with a gentle directional key, identical across all panels, with a neutral cool-gray color grade. The poses are confident and high-fashion, with attitude, never stiff.
Reading left to right and top to bottom: panel one is a dramatic worm's-eye shot from floor level, the subject towering over the lens, weight on one leg, one hand in a pocket, chin slightly lowered, gazing down coolly into the camera. Panel two is an extreme close-up cropped tightly from brow to lips, eyes locked on the lens. Panel three is a drone view from very high above, the subject small at the center of the empty gray floor, mid-stride, with a long soft shadow. Panel four is a clean side profile, chest-up, head held high, gaze fixed into the distance. Panel five is a direct top-down view looking straight down on the standing subject, who tilts the face up toward the lens, with the body foreshortened beneath. Panel six is a low three-quarter full-body shot, the subject turned three-quarters away with the head looking back over the shoulder at the camera. Panel seven is a wide full-body shot, the subject small in the empty frame, walking toward the camera mid-step. Panel eight is an over-the-shoulder shot from behind, the subject's shoulder and hair softly blurred in the foreground, face turned in three-quarter profile. Panel nine is a high-angle shot looking down, the subject sitting on the floor with one knee up and an arm resting on it, looking up into the lens.
At the bottom of each panel, place a white rounded rectangular label box with clean black sans-serif text, matching position and size in every panel, with exactly these texts in order: "/wormseye:", "/extremecloseup:", "/droneview:", "/profile:", "/topdown:", "/lowthreequarter:", "/widefull:", "/overshoulder:", "/highangle:". The labels must be clearly readable and correctly spelled.
Photorealistic, premium fashion editorial photography, natural skin texture with visible pores, realistic anatomy and hands, realistic hair and fabric detail. No other text, no logos, no watermarks, no reference-photo inset, no additional people, no background objects.Models & stability
Recommended models
Keep exploring

The author uses Nano Banana for commercial portrait retouching of an uploaded photo. The core workflow locks face identity and proportions and rebuilds only soft studio lighting, a pure white background and natural skin texture, with explicit negative constraints against changing the face.
Prompt
Preserve the person's face exactly as it appears in the uploaded image. Do NOT regenerate, reconstruct, reinterpret, beautify, reshape, or replace the face. Maintain the exact: Facial structure and

The VibeShot prompt framework from Guyi's X long post: the subject stays fixed while expression, clothing, scene, moment, shot size, focal length, camera position, composition, foreground and lighting each get a variable pool, so GPT-Image 2 randomizes within limits to produce 10 candid 9:16 shots at once.

📸 Upload your photo → paste the prompt → let AI transform it into a stunning poster. GPT Image 2.5 Prompt: Turn this portrait photo into a single high-end half…
Prompt
Turn this portrait photo into a single high-end halftone travel-collage poster. Output one poster per uploaded photo — never merge multiple photos into one frame. FORMAT 3:4 vertical canvas. Split i…
Same creator
Seedance: Surely No One Thinks This Is Real, Right? 😄AI Video · Seedance · 2026-09-06
Sweet-to-Edgy Makeover After Removing Her GlassesAI Video · MiniMax · 2026-08-31
Rainy-Night Limo Romance ClashAI Video · Seedance · 2026-10-03
Over-Water Human Wall Crab Squat ChallengeAI Video · Seedance · 2026-09-09
Hotel Bathrobe Shoe-Toss MakeoverAI Video · MiniMax H3 · 2026-09-02Each detail page keeps the original result, the prompt, and retest evidence in one reading path.