The small online shop
Each new piece is photographed flat on the table, then worn on a white background for the product page: no studio to rent, no model to pay for the day.
Put a flat lay garment on a model with AI
A photo of a garment laid flat and a photo of a model become a white background product shot and a worn photo outdoors, the garment intact. The workflow is shown here, no account needed.
Real workflow: a mustard yellow cardigan photographed flat on a wooden floor, then worn by Frank on a white background and in an autumn street. Five buttons, two pockets and the cable knit are kept.
At a glance
- You provide
- 2 photos · a text or a brief
- What the workflow produces
- 2 images in 3:4
- Canvas structure
- 7 nodes wired by 6 connections
- Models wired in
- Nano Banana Pro
- Cost of one full run
- about 38 credits, roughly $0.46
- Account
- The demo is visible without an account; generating requires a free account, trial credits included
A garment photographed flat sells less than a worn one: you see neither the fit, nor the drape, nor the length. But a model and a shoot cost too much for every piece of a small shop. This workflow starts from the flat photo and a photo of a model, and returns the white background product shot, then the worn photo outdoors.
The hard part is keeping THE garment as it goes from flat to volume. Nano Banana Pro receives both photos in the same node; the prompt names the first as the garment and the second as the model, and counts the details, buttons, pockets, knit. A named detail is a kept detail.
To add colourways afterwards, see change the colour of a garment; for a garment on a shop mannequin, the ghost mannequin. The two photos cost thirty-eight credits.
Actual renders from the demo
These are the files of the workflow above, exactly as the models produced them, unretouched. The starting point is shown when there is one.
Starting point
Render
Who it is for
Online shops, resellers, fashion brands and makers who photograph their pieces flat and want to show them worn.
Which files to provide
A flat photo of the whole garment in good light, and a front photo of the model, or a description.
How it works
- 1
Photograph the garment perfectly flat
From above, whole, smoothed out, in daylight. Buttons, pockets and texture must be readable: that is what the model copies.
- 2
Add the photo of the model
A sharp photo of the person, from the front, or a reference sheet. Without a photo, describe the model in the prompt and cut the wire.
- 3
Count the garment's details
Number of buttons, pockets, collar, length, pattern. The prompt writes them down, and the render keeps them.
- 4
Check before publishing
Count the buttons and compare the colour with the flat photo. An invented detail on a product page means a returned order.
Variations and use cases
The same canvas, tuned for different needs: each variation is two or three lines of prompt away.
Second-hand resale
A pre-loved garment shown worn gives an idea of drape and length, as long as the real photo stays in the listing too so the buyer sees its actual condition.
The brand preparing a collection
A collection's prototypes, laid flat at the office, become worn photos for the presentation lookbook shown to buyers, well before the final photo shoot.
The knitter or crocheter
The finished crochet or knitted piece, photographed flat, is shown worn to sell it online or to illustrate the page of the pattern being sold.
The supplier catalogue
A wholesaler who only receives flat photos from suppliers turns them into consistent worn visuals, ready for resellers to use in their own shops.
Social media
The worn photo outdoors, in autumn light, feeds a social media post or ad, from the same flat photo as the product page, without a new shoot.
Common mistakes
A wrinkled or skewed flat photo
The model copies what it sees: crushed folds become flaws, an angle distorts the length.
Not naming the details
Without "five buttons, two pockets", the count changes. Counting is the only guarantee.
Leaving the white background open
Ask for one single photograph with no insets: the first versions placed blurred close-ups of the knit around the model.
Using a model's photo without consent
A real person's face does not go on a product page without their permission.
Recommended models
Nano Banana Pro
Flux Kontext
GPT Image 2.5 Sunburst
Seedream 5 ProFrequently asked questions
Do I need a real model?
No: a photo of yourself, of someone who agrees, or a description in the prompt is enough. For a real person, ask for their consent.
What about womenswear or kidswear?
The same recipe works: change the model photo and describe the size you want. The garment stays the one in the flat photo.
Can I show the back?
Yes, with a flat photo of the back and a prompt that says "back view". Each view needs its own photo of the garment.
How much do the two photos cost?
Thirty-eight credits, nineteen per render on Nano Banana Pro. A third view costs nineteen more.
Why Nano Banana Pro?
It holds two photos at once, the garment and the face, when the prompt names each by its rank. Measured on the demo and on the group photo template.
Is the garment faithful?
In the demo, yes: same yellow, five buttons, two pockets, cable knit. Still count the details before publishing.
Can I change the pose?
Yes, in the prompt: three quarter, hands in pockets, seated. Keep a simple pose on the product page, where the garment must read clearly.
What about a dress or trousers?
The same recipe works for any piece: describe the length, the cut and the details visible on the flat photo.
Put a flat lay garment on a model with AI
Use this workflowFree account, no card required. The workflow opens pre-filled in your canvas.