1 take, 6 roupas: o workflow Seedance 2.5 Edit + Claude (com os prompts)

Photorealistic e-commerce flat-lay product photo, 3:4 vertical, of a complete men's formal outfit laid out neatly on a seamless light-grey studio background, camera directly overhead, nothing else in frame. THE OUTFIT: a classic slim-fit two-piece business suit in deep navy blue fine wool (single-breasted jacket with two buttons, notch lapels, flap pockets; matching flat-front trousers), a crisp white dress shirt with a point collar, a solid burgundy silk necktie tied in a neat knot, and a pair of black polished leather oxford shoes below the trousers. Clean, sharp, real fabric texture, soft even studio light, soft contact shadows. No people, no mannequin, no hangers, no logos, no text, no labels.

V2V - change only the clothes, nothing else. The result is the reference video re-dressed: in every frame the camera, framing, room, light, the armchair, the ottoman and the remote on it, his face, glasses, eyes, eyebrows, beard, hair, skin, expressions, lip sync, every movement and its timing are the same as in the same frame of the reference video. Frame n of the result = frame n of the reference video; second N = second N. If any sentence conflicts with the reference video, the reference video wins. Source roles: - The reference video: the only source for everything except the clothes - the camera, him (face, rectangular dark-framed glasses, short dark beard, dark hair) and his whole performance, the room, light and sound. - The reference image: take only the clothes (the navy suit, white dress shirt, burgundy tie and black oxford shoes; not the extra folded shirt) - style, color, fabric, every detail. Do not take its grey background or the flat-lay layout. RULE 0 - performance 1:1: nothing retimed, sped up, slowed down, frozen or re-acted. Every hand gesture, every finger, arm, head turn, posture, leg position at the same moment as in the source. His mouth moves exactly as in the source, in sync with his original voice. Same man, same face shape and proportions, not thinner, not heavier, not beautified; his glasses stay on his face in every frame. PERFORMANCE TRACK: 0-1 s he is still dropping into the armchair, leaning in close to the camera, exactly as in the source; from 1 s he is seated and talks with his hands. He already wears the suit while sitting down. THE NEW OUTFIT, worn from the first frame to the last, replacing his black overshirt, black pants and white sneakers: a deep navy blue slim-fit two-piece wool suit (single-breasted jacket worn on, notch lapels), a crisp white point-collar dress shirt, a burgundy silk tie in a neat knot, navy suit trousers and black polished leather oxford shoes. Worn naturally like his own clothes, not a costume; it folds and moves with every real movement, lit by the room's daylight with soft contact shadows; never pasted on. THE ROOM stays identical: locked-off static camera; the orange-tan wooden armchair, the orange ottoman with the black remote in the foreground, the big fiddle-leaf plant, the colorful abstract paintings, the lilac and grey curtains, the light grey floor. FAILURE: glasses missing or changed; face changed, wider, heavier, darker beard, or turned into someone else; skin plastic or smoothed; mouth out of sync; gesture missing, early, late, mirrored or with the other hand; hands frozen where the source moves; extra fingers; the original black clothes showing; outfit wrong, incomplete or changing over time; posture changed; armchair, ottoman, remote, plant, paintings or curtains changed; camera move, zoom or cut; extra people; text, logos, subtitles, music. Sound: keep the original audio exactly as it is.

V2V - change only the clothes, nothing else. The result is the reference video re-dressed: in every frame the camera, framing, room, light, the armchair, the ottoman and the remote on it, his face, glasses, eyes, eyebrows, beard, hair, skin, expressions, lip sync, every movement and its timing are the same as in the same frame of the reference video. Frame n of the result = frame n of the reference video; second N = second N. If any sentence conflicts with the reference video, the reference video wins. Source roles: - The reference video: the only source for everything except the clothes - the camera, him (face, rectangular dark-framed glasses, short dark beard, dark hair) and his whole performance, the room, light and sound. - The reference image: take only the clothes (the long white lab coat, light-blue scrub top, light-blue scrub pants and white sneakers) - style, color, fabric, every detail. Do not take its grey background or the flat-lay layout. RULE 0 - performance 1:1: nothing retimed, sped up, slowed down, frozen or re-acted. Every hand gesture, every finger, arm, head turn, posture, leg position at the same moment as in the source. His mouth moves exactly as in the source, in sync with his original voice. Same man, same face shape and proportions, not thinner, not heavier, not beautified; his glasses stay on his face in every frame. THE NEW OUTFIT, worn from the first frame to the last, replacing his black overshirt, black pants and white sneakers: a long white cotton dentist lab coat worn open (long sleeves, notch collar, white buttons, patch pockets, a pen in the chest pocket) over a light-blue crew-neck scrub top, light-blue scrub pants and white sneakers. No stethoscope, no mask. Worn naturally like his own clothes, not a costume; it folds and moves with every real movement, lit by the room's daylight with soft contact shadows; never pasted on. THE ROOM stays identical: locked-off static camera; the orange-tan wooden armchair, the orange ottoman with the black remote in the foreground, the big fiddle-leaf plant, the colorful abstract paintings, the lilac and grey curtains, the light grey floor. FAILURE: glasses missing or changed; face changed, wider, heavier, darker beard, or turned into someone else; skin plastic or smoothed; mouth out of sync; gesture missing, early, late, mirrored or with the other hand; hands frozen where the source moves; extra fingers; the original black clothes showing; outfit wrong, incomplete or changing over time; posture changed; armchair, ottoman, remote, plant, paintings or curtains changed; camera move, zoom or cut; extra people; text, logos, subtitles, music. Sound: keep the original audio exactly as it is.

V2V - change only the clothes, nothing else. The result is the reference video re-dressed: in every frame the camera, framing, room, light, the armchair, the ottoman and the remote on it, his face, glasses, eyes, eyebrows, beard, hair, skin, expressions, lip sync, every movement and its timing are the same as in the same frame of the reference video. Frame n of the result = frame n of the reference video; second N = second N. If any sentence conflicts with the reference video, the reference video wins. Source roles: - The reference video: the only source for everything except the clothes - the camera, him (face, rectangular dark-framed glasses, short dark beard, dark hair) and his whole performance, the room, light and sound. - The reference image: take only the clothes (the teal-green surgical scrub top, scrub pants, scrub cap and white clogs; not the face mask) - style, color, fabric, every detail. Do not take its grey background or the flat-lay layout. RULE 0 - performance 1:1: nothing retimed, sped up, slowed down, frozen or re-acted. Every hand gesture, every finger, arm, head turn, posture, leg position at the same moment as in the source. His mouth moves exactly as in the source, in sync with his original voice. Same man, same face shape and proportions, not thinner, not heavier, not beautified; his glasses stay on his face in every frame. THE NEW OUTFIT, worn from the first frame to the last, replacing his black overshirt, black pants and white sneakers: a teal-green (hospital green) short-sleeve V-neck surgical scrub top, matching teal-green drawstring scrub pants, a matching teal-green cloth surgeon's scrub cap on his head covering the top of his hair (his glasses, forehead and whole face stay visible), and white clogs. NO face mask. Worn naturally like his own clothes, not a costume; it folds and moves with every real movement, lit by the room's daylight with soft contact shadows; never pasted on. THE ROOM stays identical: locked-off static camera; the orange-tan wooden armchair, the orange ottoman with the black remote in the foreground, the big fiddle-leaf plant, the colorful abstract paintings, the lilac and grey curtains, the light grey floor. FAILURE: glasses missing or changed; face changed, wider, heavier, darker beard, or turned into someone else; skin plastic or smoothed; mouth out of sync; gesture missing, early, late, mirrored or with the other hand; hands frozen where the source moves; extra fingers; the original black clothes showing; outfit wrong, incomplete or changing over time; posture changed; armchair, ottoman, remote, plant, paintings or curtains changed; camera move, zoom or cut; extra people; text, logos, subtitles, music. Sound: keep the original audio exactly as it is.

V2V - change only the clothes, nothing else. The result is the reference video re-dressed: in every frame the camera, framing, room, light, the armchair, the ottoman and the remote on it, his face, glasses, eyes, eyebrows, beard, hair, skin, expressions, lip sync, every movement and its timing are the same as in the same frame of the reference video. Frame n of the result = frame n of the reference video; second N = second N. If any sentence conflicts with the reference video, the reference video wins. Source roles: - The reference video: the only source for everything except the clothes - the camera, him (face, rectangular dark-framed glasses, short dark beard, dark hair) and his whole performance, the room, light and sound. - The reference image: take only the clothes (the heather-grey tank top, black shorts, white socks and black-and-white running sneakers) - style, color, fabric, every detail. Do not take its grey background or the flat-lay layout. RULE 0 - performance 1:1: nothing retimed, sped up, slowed down, frozen or re-acted. Every hand gesture, every finger, arm, head turn, posture, leg position at the same moment as in the source. His mouth moves exactly as in the source, in sync with his original voice. Same man, same face shape and proportions, not thinner, not heavier, not beautified; his glasses stay on his face in every frame. THE NEW OUTFIT, worn from the first frame to the last, replacing his black overshirt, black pants and white sneakers: a heather-grey sleeveless dry-fit athletic tank top (his own arms and shoulders bare, same slim build, not more muscular), black athletic shorts above the knee with a grey waistband (his own lower legs bare), white crew sports socks and black-and-white running sneakers. Worn naturally like his own clothes, not a costume; it folds and moves with every real movement, lit by the room's daylight with soft contact shadows; never pasted on. THE ROOM stays identical: locked-off static camera; the orange-tan wooden armchair, the orange ottoman with the black remote in the foreground, the big fiddle-leaf plant, the colorful abstract paintings, the lilac and grey curtains, the light grey floor. FAILURE: glasses missing or changed; face changed, wider, heavier, darker beard, or turned into someone else; skin plastic or smoothed; mouth out of sync; gesture missing, early, late, mirrored or with the other hand; hands frozen where the source moves; extra fingers; the original black clothes showing; outfit wrong, incomplete or changing over time; posture changed; armchair, ottoman, remote, plant, paintings or curtains changed; camera move, zoom or cut; extra people; text, logos, subtitles, music. Sound: keep the original audio exactly as it is.

V2V - change only the clothes, nothing else. The result is the reference video re-dressed: in every frame the camera, framing, room, light, the armchair, the ottoman and the remote on it, his face, glasses, eyes, eyebrows, beard, hair, skin, expressions, lip sync, every movement and its timing are the same as in the same frame of the reference video. Frame n of the result = frame n of the reference video; second N = second N. If any sentence conflicts with the reference video, the reference video wins. Source roles: - The reference video: the only source for everything except the clothes - the camera, him (face, rectangular dark-framed glasses, short dark beard, dark hair) and his whole performance, the room, light and sound. - The reference image: take only the clothes (the turquoise Hawaiian floral short-sleeve shirt, beige shorts and sandals) - style, color, fabric, every detail. Do not take its grey background or the flat-lay layout. RULE 0 - performance 1:1: nothing retimed, sped up, slowed down, frozen or re-acted. Every hand gesture, every finger, arm, head turn, posture, leg position at the same moment as in the source. His mouth moves exactly as in the source, in sync with his original voice. Same man, same face shape and proportions, not thinner, not heavier, not beautified; his glasses stay on his face in every frame. THE NEW OUTFIT, worn from the first frame to the last, replacing his black overshirt, black pants and white sneakers: a turquoise Hawaiian shirt with a tropical floral print, short sleeves, worn open-collar (his own forearms bare), beige chino shorts above the knee (his own lower legs bare) and brown leather sandals. Worn naturally like his own clothes, not a costume; it folds and moves with every real movement, lit by the room's daylight with soft contact shadows; never pasted on. THE ROOM stays identical: locked-off static camera; the orange-tan wooden armchair, the orange ottoman with the black remote in the foreground, the big fiddle-leaf plant, the colorful abstract paintings, the lilac and grey curtains, the light grey floor. FAILURE: glasses missing or changed; face changed, wider, heavier, darker beard, or turned into someone else; skin plastic or smoothed; mouth out of sync; gesture missing, early, late, mirrored or with the other hand; hands frozen where the source moves; extra fingers; the original black clothes showing; outfit wrong, incomplete or changing over time; posture changed; armchair, ottoman, remote, plant, paintings or curtains changed; camera move, zoom or cut; extra people; text, logos, subtitles, music. Sound: keep the original audio exactly as it is.

V2V - change only the clothes, nothing else. The result is the reference video re-dressed: in every frame the camera, framing, room, light, the armchair, the ottoman and the remote on it, his face, glasses, eyes, eyebrows, beard, hair, skin, expressions, lip sync, every movement and its timing are the same as in the same frame of the reference video. Frame n of the result = frame n of the reference video; second N = second N. If any sentence conflicts with the reference video, the reference video wins. Source roles: - The reference video: the only source for everything except the clothes - the camera, him (face, rectangular dark-framed glasses, short dark beard, dark hair) and his whole performance, the room, light and sound. - The reference image: take only the clothes (the old faded heather-grey t-shirt, navy sweat shorts and black slide sandals) - style, color, fabric, every detail. Do not take its grey background or the flat-lay layout. RULE 0 - performance 1:1: nothing retimed, sped up, slowed down, frozen or re-acted. Every hand gesture, every finger, arm, head turn, posture, leg position at the same moment as in the source. His mouth moves exactly as in the source, in sync with his original voice. Same man, same face shape and proportions, not thinner, not heavier, not beautified; his glasses stay on his face in every frame. THE NEW OUTFIT, worn from the first frame to the last, replacing his black overshirt, black pants and white sneakers: an old, faded, slightly stretched heather-grey crew-neck cotton t-shirt with a worn collar (his own forearms bare), loose navy-blue cotton drawstring sweat shorts above the knee (his own lower legs bare) and black rubber slide sandals, bare feet. Casual stay-at-home clothes, just-woke-up look. Worn naturally like his own clothes, not a costume; it folds and moves with every real movement, lit by the room's daylight with soft contact shadows; never pasted on. THE ROOM stays identical: locked-off static camera; the orange-tan wooden armchair, the orange ottoman with the black remote in the foreground, the big fiddle-leaf plant, the colorful abstract paintings, the lilac and grey curtains, the light grey floor. FAILURE: glasses missing or changed; face changed, wider, heavier, darker beard, or turned into someone else; skin plastic or smoothed; mouth out of sync; gesture missing, early, late, mirrored or with the other hand; hands frozen where the source moves; extra fingers; the original black clothes showing; outfit wrong, incomplete or changing over time; posture changed; armchair, ottoman, remote, plant, paintings or curtains changed; camera move, zoom or cut; extra people; text, logos, subtitles, music. Sound: keep the original audio exactly as it is.