products and the person wears all of them in one result: a shirt with trousers, a dress under a trench coat, a full suit. Only those garments change. Everything else the person wears, shoes included, stays as it is in their photo.
Works with
POST /try-on, POST /try-on/sync and batches. Every item takes the same shapes as a single product: a reference, an inline product, or a one-shot.
Example
products does not matter for layering: jackets, coats and vests are worn over the tops automatically. The response lists every product in productExternalIds, in the order you sent them.
Best results
- Send the title, and the description when you have it. They tell the AI which garment of the photo is the product, especially on photos where the model wears a full outfit.
- Packshots work best, photos on a model work well too. Use the photo that shows the whole garment.
- One item per part of the body, plus outer layers: a top and a bottom, or a dress, then a jacket or coat on top. Two bottoms in one try-on give unpredictable results.
- Use a full-body photo of the person when the look includes trousers, skirts or a long coat. On a half-body photo, only what fits in the frame is visible.
Limits
- Clothing only. A try-on with glasses, a hat, shoes, a bag or jewellery among several products fails with
MULTI_PRODUCT_NOT_SUPPORTEDand the credits are refunded. Send those products in their own try-on. - Up to 4 products. A fifth returns
400 VALIDATION_FAILED. - Uploaded product files share the request’s 10-file limit.

