Get started
Rate limits
Was this page helpful?
Design clients that stay inside request limits and recover cleanly from HTTP 429 responses.
Current limits
| Request | Limit |
|---|---|
| List models and contract reads | 60/minute per trusted client address |
| Generate media | 10/minute and 200/day per key |
| Video account cap | 10 per hour and 100 per day per account until the account has a paid history; 200 per day after |
| Job status | 120/minute per key |
Retry-After
HTTP 429 includes Retry-After. Pause for at least that many seconds, add jitter, and do not retry concurrent copies of the same request.
Backoff pattern
For polling, start around 2 seconds and increase to 5–10 seconds. Stop when the job reaches a terminal state. For generation retries, reuse the same Idempotency-Key and body.