Rate limits & fair use
Current limits and resilient client behavior.
Current policy
The API allows 300 requests per minute per IP. RateLimit-Remaining and RateLimit-Reset response headers expose the current budget. Email signup links also have a recipient-specific daily limit. Large imports and report generation are asynchronous and should be polled conservatively.
Design for backpressure
Honor 429 responses, use exponential backoff with jitter, and poll asynchronous jobs every 5–10 seconds rather than in a tight loop.
Recommended retry policy
| Response | Retry? | Guidance |
|---|---|---|
429 | Yes | Use Retry-After when present; otherwise exponential backoff. |
5xx | Yes | Retry idempotent reads and safe writes with jitter. |
4xx except 429 | No | Correct the request before retrying. |