VODOOAI lookbook: one model, one palette, a whole collection in a day
By VODOO AI · · 4 min read
A lookbook works because it feels like one day: the same person, the same place, the same light, the collection moving through it. That's exactly what AI finds hardest by default, because every image starts fresh.
So an AI lookbook is mostly about holding things still. Fix the model, the palette, the location and the light in words you reuse every frame, and change only the garment and the pose.
Write the shoot's bible first
Before any garment, write the paragraph that never changes. It describes the model like a casting card, the location, the time of day, the lens and the colour. Every frame starts with it, word for word.
Model: a woman in her late twenties, East African, close-cropped natural hair, freckles across the nose, calm expression, natural skin texture with visible pores. Location: a sun-bleached stone courtyard in southern Spain with whitewashed walls and a terracotta floor. Light: late afternoon sun from the left, warm, long soft shadows. Camera: 50mm, full length, eye level. Colour: warm, muted, slight film grain.Make this in VODOO →
The frame list
A lookbook is usually six to twelve frames per drop. Plan them like a photographer would: a mix of full length, three-quarter and detail, a mix of still and walking, and one or two that show the place.
1
- Shot
- Full length, standing
- What it shows
- The whole outfit, clearly
2
- Shot
- Walking towards camera
- What it shows
- How the fabric moves
3
- Shot
- Three-quarter, seated on a step
- What it shows
- Fit through the body
4
- Shot
- Detail: hands, cuff, texture
- What it shows
- Quality and material
5
- Shot
- Wide: model small in the courtyard
- What it shows
- The world of the collection
6
- Shot
- Over the shoulder, looking away
- What it shows
- Mood, for the closing page
| Frame | Shot | What it shows |
|---|---|---|
| 1 | Full length, standing | The whole outfit, clearly |
| 2 | Walking towards camera | How the fabric moves |
| 3 | Three-quarter, seated on a step | Fit through the body |
| 4 | Detail: hands, cuff, texture | Quality and material |
| 5 | Wide: model small in the courtyard | The world of the collection |
| 6 | Over the shoulder, looking away | Mood, for the closing page |
Describe the garment like a buyer would
The engine can only render what you name. Say the fabric, the weight, the colour, the cut and the details, and keep that description the same every time the piece appears.
[Bible paragraph.] She walks towards camera across the courtyard wearing a loose ivory linen shirt dress, mid-calf, rolled sleeves, a thin tan leather belt, flat tan sandals. The linen creases and moves with each step; the hem catches the breeze.Make this in VODOO →
[Bible paragraph, camera changed to 100mm close-up.] Her hand rests on the rolled linen sleeve of the ivory shirt dress, showing the slub texture of the fabric and the stitched cuff, the thin tan belt just in frame. Sun on the back of her hand, visible knuckle creases.Make this in VODOO →
When the model drifts
Even with the same words, a face can change between frames. Pick the best frame as the reference and make the rest from it: in VODOO, add that picture to the brief in Create and say "same model as in the picture". Nano Banana and Seedream are good at keeping a person consistent across a set, and Flux gives a rich photographic finish.
To make the walking frames move, Kling handles fabric and stride well. Our longer guide to AI fashion models covers on-model shots and video for every colourway.
Being honest about it
Use invented models or real people who've agreed in writing to exactly this use. Label AI imagery where it's required, and keep a real product photo for close fit and fabric questions. Start your first lookbook with 300 free credits a month.
Questions people ask
- Can AI make a fashion lookbook?
- Yes. Write one paragraph describing the model, location, light and camera, reuse it word for word on every frame, and change only the garment and pose. Use your best frame as a reference to keep the model consistent.
- How do I keep the same model across AI images?
- Use the same casting description every time, then make further frames from your best image as a reference picture. Some engines are noticeably better at keeping a person consistent across a set.
- How many images should a lookbook have?
- Six to twelve frames per drop is typical: a mix of full length, three-quarter, detail and one or two wide shots that show the place.