Rate Limiting
Rate limiting is enforced at workspace level. Different keys belonging to the same workspace will share the same quotas. By default, each endpoint has a limit of 3 calls per second, with some allowance for burstiness. A global limit of 3000 calls per day is enforced at workspace level. When one of these limits is reached, the request will fail with a429 Too many requests error.
Information about the current quotas can be retrieved from the response headers:
- X-RateLimit-Limit: The sustained request rate for this endpoint, in requests per second. This is how quickly your rate limit quota refills.
- X-RateLimit-Remaining: The number of requests you can make immediately before being rate-limited. This value can exceed X-RateLimit-Limit because the API allows short bursts of requests. Once depleted, it refills at the rate shown in X-RateLimit-Limit.
- X-RateLimit-Daily-Remaining: The number of API calls remaining for your workspace within a rolling 24-hour window. Requests made more than 24 hours ago no longer count against this limit.
