master defaults, the window is 60 seconds, with 120 requests per API key and 600 requests per IP. Deployment configuration can override these values, so response headers remain authoritative.
Default behavior
Rate limits are evaluated by IP address and API key. When a request is allowed, the response includes rate-limit headers for the active API-key bucket when available.X-RateLimit-Reset is a Unix timestamp in seconds.
Limit exceeded
When a request exceeds the active limit, the API returns429 Too Many Requests.
Retry-After seconds before retrying.
Integration guidance
- Use pagination instead of large repeated requests.
- Back off when
X-RateLimit-Remainingis low. - Retry
429responses only afterRetry-After. - Include the
traceIdwhen contacting support about rate-limit behavior. - In Clay or another scheduler, start at no more than
120requests per60000ms for one API key and lower the rate when several tables or workers share that key.