Eleven workflows, one model
try-on, extract, look, the model / background / garment
generators, and the swap ops are peers over one job surface. Learn one
lifecycle, call any of them.
Figur API
An AI fashion-imagery engine. Dispatch a workflow, poll the event stream, fetch the render — eleven workflows, one job model.
One engine, one model, one bill
Section titled “One engine, one model, one bill”Figur turns product and model imagery into finished fashion renders. Try-on, extraction, look composites, and on-model / background / garment generation are all peers over a single job model — so you learn one request shape and one lifecycle, then call any workflow the same way.
Dispatch → events → render
Every call returns a job_id immediately, streams
queued → stage → succeeded events, and resolves to a render —
identical across all workflows.
Credits, transparently
Flat per-dispatch pricing, deducted synchronously and refunded automatically when a render fails. Live costs come from the API, never hard-coded.
Generated, never stale
This reference is generated straight from the gateway’s OpenAPI document, so it can’t drift from the running code.
Start here
Section titled “Start here” Getting started The dispatch → poll → fetch lifecycle, with curl and TypeScript SDK examples.
Authentication Bearer API keys and the jobs:create / jobs:read capabilities.
Errors & credits The error envelope, per-op credit costs, refund-on-failure, and rate limits.
API reference Every endpoint, parameter, and schema — generated from openapi.json.