How the window works
- Each limit uses a fixed 60-second window. The window starts with your first request in that group and resets 60 seconds later; it does not slide.
- Every endpoint group has its own counter. Requests to
/v1/api/paymentdon’t use up your/v1/api/payoutbudget. - Some groups have a second, tighter limit on top of the group limit. A request must pass every limit that applies to it.
- Requests that are rejected with
429still count towards the current window.
Limits
Rate limit headers
Responses from rate-limited endpoints include these headers:
Where more than one limit applies (for example, on payout routes), the headers describe the tighter limit.
When you exceed a limit
The API responds with HTTP429 and the standard error envelope. The message states the limit you hit:
- Stop sending requests to that group.
- Wait for
RateLimit-Resetseconds. - Retry. For payouts, retry with the same
Idempotency-Key; see Idempotency.