Rate limits
Per key, per minute, over a rolling 60-second window. The limit depends on your plan.
Each API key has a per-minute rate limit, enforced server-side over a rolling window. Every request is metered per key. The caller's live budget is returned on every authenticated response, so a client never has to guess how much headroom it has.
| Plan | Requests / minute | Requests / month |
|---|---|---|
| Free | 10 | 1,000 |
| Pro | 120 | 100,000 |
| Business | 300 | 500,000 |
| Enterprise | custom (600+) | 2,000,000+ |
Budget headers
Every response, including 200, carries the current budget:
RateLimit-Limit: 120 RateLimit-Remaining: 87 RateLimit-Reset: 23 RateLimit-Policy: 120;w=60 X-RateLimit-Limit: 120 X-RateLimit-Remaining: 87 X-RateLimit-Reset: 23 X-Quota-Limit: 100000 X-Quota-Remaining: 99913
RateLimit-* follow the IETF rate-limit header draft; the X-RateLimit-* names are the legacy equivalents, kept for compatibility. RateLimit-Reset is the number of seconds until the window frees up. X-Quota-* report the monthly quota and what is left of it.
Exceeding the per-minute limit
HTTP/1.1 429 Too Many Requests Retry-After: 23 RateLimit-Limit: 120 RateLimit-Remaining: 0 RateLimit-Reset: 23
Exceeding the monthly quota
HTTP/1.1 429 Too Many Requests
Retry-After: 3600
X-Quota-Limit: 100000
X-Quota-Remaining: 0
{"detail":"monthly quota exceeded"}
RateLimit-Remaining and back off before you hit zero; on a 429, wait Retry-After seconds. Cache responses where possible, and bulk consumers should batch by since instead of polling continuously. See Pricing for the full plan comparison.