Information
- OpenAPI version:
3.0.3
A client’s own server or AI agent, holding a key for ONE workspace, uses this API to run
restaurant video production. Every response carries a Request-Id header; every error has
the envelope {error:{type, code, message, field?, doc_url, request_id}}.
Rate limit: 600 requests a minute per workspace, across every server that answers. Above it,
429 rate_limited with Retry-After. Twenty failed authentications (missing or unknown key)
from one address within a minute block that address for 15 minutes.
Authorization: Bearer crp_live_… or crp_rk_live_… (live), crp_test_… or crp_rk_test_… (sandbox).
Security scheme type: http
Authorization: Bearer crp_obs_… (the session_secret from POST /onboarding_sessions). Polls one session only.
Security scheme type: http