Rate limits
Engine ≥ 4.143.0. Each token response includes data.rate_limits from the store policy at issuance. All keys share these budgets; a new token does not reset them. A plan change may apply before the token expires.
- read_per_minute
- Read requests per minute.
- write_per_minute
- Write requests per minute; this value does not grant write permissions.
- auth_per_minute
- Token requests per minute, including failed authentication attempts.
Send requests evenly over time; for a positive read limit, use at least 60 / read_per_minute seconds between reads. The limit is not a promise of a simultaneous burst. Other integrations share it: handle 429 and Retry-After.
Zero means unavailable, not unlimited.
Limits per minute
| Available from | Read · GET |
|---|---|
| BASIC | 60 |
| STANDARD | 120 |
| PROFESSIONAL | 300 |
| ADVANCED | 900 |
- Write
- 10
- Authentication
- 10
These limits are shared by all keys belonging to the same store.
Request
- Read and write traffic is subject to the store's current API policy.
- Keys do not provide separate unlimited capacity.
- Keep concurrency bounded, cache reference data where appropriate and avoid repeatedly fetching unchanged detail records.
429 · Retry-After
- When the API returns 429 with rate_limit.exceeded, wait at least the number of seconds in Retry-After and add jitter before retrying.
- 429 · capacity.queue_full — The short API queue is full; the request has not started execution. 429 · capacity.wait_timeout — The API queue wait expired; the request has not started execution. Wait at least Retry-After seconds and add jitter. Preserve the original key and payload when retrying a write.
- Use the documented error code and header instead of guessing a fixed retry interval.