Rate limiting

The API is limited to 60 requests per minute per account by default. Exceeding the limit returns `429 Too Many Requests`. Requests may occasionally be throttled with the same status code during periods of high load.

Rate-limit headers

Every response includes headers describing your current rate-limit window so you can pace requests without guessing.

Response headers
ParameterTypeDescription
X-RateLimit-LimitintegerMaximum requests allowed in the current window.
X-RateLimit-RemainingintegerRequests remaining in the current window.
X-RateLimit-ResetintegerUnix timestamp (seconds) when the window resets.
Retry-AfterintegerOn a `429`, the number of seconds to wait before retrying.

Handling 429 responses

  1. When you receive a `429`, read the `Retry-After` header and pause for that many seconds.
  2. Retry the request with exponential backoff (for example, 1s, 2s, 4s) plus a little random jitter.
  3. Keep your steady-state request rate under the limit by spacing calls or queueing work.
  4. For large batch jobs, prefer endpoints that accept multiple inputs per request over many single-item calls.
Design for idempotent retries

Read endpoints (GET) are safe to retry. For metered operations, a `429` means the request was rejected before any credits were charged - retrying does not double-charge.