Rate Limits
API endpoints are rate limited to protect the platform from abusive traffic without slowing normal work.
How requests are counted
Section titled “How requests are counted”- A signed-in request counts against your company, so everyone on one company shares a limit. A signed-in request with no company counts against the person. A request with no valid token counts against the caller’s IP address.
- The default is 100 requests a minute.
- Some routes are tighter. Login allows 5 a minute, and signup, phone verification and the routes that start AI work have lower limits still.
- Public API keys have their own limit per key, 100 an hour by default.
- Health checks are not limited.
These are the service’s defaults and can change. Do not build around a number. Build around the response.
The response
Section titled “The response”When you exceed a limit, the API returns:
429 Too Many Requestswith a JSON body, {"detail": "Rate limit exceeded", "limit": "..."}, and a Retry-After header telling you how many seconds to wait before retrying. Normal responses carry no remaining-requests header, so the 429 is the only signal.
Working with limits
Section titled “Working with limits”- Space out batch operations instead of firing them all at once.
- Respect
Retry-Afterwhen it appears. Retrying immediately keeps you limited. - If a legitimate workflow needs a higher limit, talk to your account contact rather than working around the limiter.