Skip to main content
The base URL is https://api.bitstudio.ai. Authenticate with a Studio workspace API key; requests use that workspace’s access and credits.
  1. Upload images or select saved assets.
  2. Submit directly to the operation you want, such as POST /v1/images/virtual-try-on or POST /v1/images/{id}/edit.
  3. Save the returned job IDs and poll GET /v1/jobs/{id}. On completion, use result.url and result.id.
A job tracks the progress of your request. You create it through the operation endpoint; every creation operation returns the same response shape.

Try on one item

Put a garment on a specific person with V6.

Build a complete outfit

Combine up to four garments in one result.

Generate a product photo

Create a new fashion image using your product references.

Edit an image

Change the background, pose, or details of a completed image.

​
A consistent contract

Creation requests use plain operation fields: model, resolution, and the relevant references. Try-on defaults to v6; generate and text edit default to nano-banana-2. Omitted image counts default to one. Unknown input fields are rejected, and optional creation settings should be omitted rather than set to null. Output IDs are opaque. Image results can be used as inputs for the next operation without understanding Studio’s version history. Treat new response fields as additive and ignore those your client does not use. Use idempotency keys to retry creation requests safely. Existing unversioned image and asset endpoints remain available under Legacy API with their existing behaviour, including the unversioned try-on model default of v1. For a shopper-facing fitting room, see the virtual.fit integration docs.