Fashion lookbook prompts: same model, same outfit, every shot
A lookbook clip has one job: show how a garment looks and moves on a body. With reference images of the model wearing the outfit, the prompt only needs to stage the walk, the turn or the gust of wind. These prompts cover studio, location and motion shots, with notes on the details that break most often, such as hems, shoes and patterns.
What makes lookbook prompts work
Shoot references of the model in the exact outfit: full body front, three-quarter and back, plus two close-ups of the details buyers check most, such as the fabric and the shoes. RefVid takes five images per subject, so choose those close-ups carefully.
Use one outfit per reference set. If you want three looks, save three characters, one per look, with the same face, rather than describing the change in the prompt.
Describe how the fabric should behave, not what it looks like. "The wide trousers swing with each step" tells the model about weight and drape; the color and cut already come from the images.
Use full-body framing for at least one shot per look. Framing that crops at the knee hides the very details buyers check: hem length, shoes and fit.
The examples call the models Elise and Noah; use the names you saved yours under. Clip length and aspect ratio are settings, shown here in brackets.
Studio lookbook shots
Prompt (6 s, 9:16): Elise walks toward the camera on a plain off-white studio floor, a relaxed runway pace, arms loose. The camera is static at hip height with a long lens, full body in frame from head to shoes. Large soft key light from the front left, gentle shadow on the floor. She stops on a mark and holds still for the final second.
Why it works: a walk toward a long lens shows fit and movement with little distortion. If the face drifts at the end, swap one detail close-up in the references for a close-up portrait.
Prompt (8 s, 9:16): Elise stands on a mark in a pale grey studio and turns slowly in place through a full circle, pausing briefly at the side and the back. Static camera, full body. Even, shadowless light from two large softboxes.
Why it works: this is the video version of a front, side and back flat lay. It only holds if the back of the outfit is in the references; otherwise the model invents it.
Prompt (5 s, 9:16): Close-up of Elise's hand running slowly down the front of the coat, fingers brushing over the fabric texture and stopping at the top button. Shallow depth of field, raking side light that shows the weave.
Why it works: raking light turns texture into shadow and highlight that the camera can read. Use a fabric close-up as a reference or the weave will be generic.
Location editorial shots
Prompt (8 s, 9:16): Elise stands at the edge of a concrete rooftop at golden hour, city blurred behind her. A steady breeze lifts the hem of the dress and the loose ends of her hair. She looks off to the left, then back to the lens. Camera slowly arcs from her right side to the front. Warm low sun as a backlight, soft bounce on the face.
Why it works: wind gives fabric a reason to move. Keep the breeze steady; a sudden gust often reshapes the garment.
Prompt (7 s, 16:9): Noah descends a wide stone staircase in a quiet plaza, one hand trailing the rail. The camera tracks backward ahead of him at chest height. Overcast midday light, muted colors, light footsteps on stone.
Why it works: stairs make trousers and coats fold and swing in a way flat walking does not. If the steps warp, pick a shorter flight.
Motion that shows the garment
Prompt (5 s, 9:16): Elise does a single slow spin on a polished wooden floor, the pleated skirt flaring out and settling as she stops facing the camera. Static camera at waist height, full body. Warm tungsten light from one side.
Why it works: one spin is a clear start and end for the fabric. Two or more spins tend to multiply pleats or change their count.
Prompt (6 s, 9:16): Noah sits down on a simple wooden chair, crosses one leg over the other and settles back, coat falling open. Medium-wide shot, static camera, soft window light.
Why it works: sitting shows how a garment creases and drapes at rest. Keep the chair plain so it does not merge with the clothes.
Building a lookbook sequence
RefVid can put up to three shots in one clip: write each shot in its own box, for example a full-body walk toward the camera, then a close-up of the coat buttons. The clip's length is split evenly across the shots, so a 6-second clip gives two 3-second shots. That is a quick way to get a wide and a detail of the same look in one generation.
"Next shot" starts a new clip from the last frame of the previous one, which helps keep the location and light continuous across a longer sequence. Pick one aspect ratio per sequence: 9:16 for social feeds, 16:9 for site banners, 1:1 for product grids.
Save each look as its own character with the same face photos, and a season's lookbook becomes a set of characters you reuse for every new scene.
Frequently asked questions
- Can I change the outfit in the prompt and keep the same model?
- It works poorly, because the reference images define both the person and the clothes. Create a new reference set for each outfit with the same model, keep the face photos the same, and save each set as its own character.
- Why do stripes and checks look wrong in AI fashion video?
- Fine, high-contrast patterns shimmer or shift as fabric moves. Add a flat close-up of the pattern to the references and prefer slower motion for those pieces.
- Which aspect ratio is best for a lookbook video?
- 9:16 for social feeds and full-body framing, 16:9 for site banners and location shots. Pick one per sequence so the cuts match.