API reference
Rate limits
Requests are rate limited per account. Exceeding a limit returns 429 with a Retry-After header (seconds).
| Scope | Limit |
|---|---|
| Per account (all keys) | See Dashboard → Billing for your account's limit. Task polling is allowed at a higher rate than creating. |
| Tasks running at once | A create call can also be refused while too many of your earlier tasks are still generating — the same 429 with Retry-After. Wait for that delay instead of adding more parallel calls: a slow model frees a slot only when a task finishes. |
Polling GET /api/v1/task is light; the limits above apply to creating tasks. Back off on 429 rather than retrying immediately.