Rate limits
Limits protect the shared atlas and keep latency low for everyone. They are applied per API key.
Default limits
| Window | Demo key | Live key |
|---|---|---|
| Per minute | 60 | 600 |
| Per day | 1,000 | 100,000 |
Rate-limit headers
Every response includes your current budget:
Response headers
X-RateLimit-Limit: 60
X-RateLimit-Remaining: 58
X-RateLimit-Reset: 1748371200When you hit the limit
Exceeding a window returns 429 with code rate_limited. Honor the X-RateLimit-Reset timestamp and retry with exponential backoff.
Limits are not enforced in this mock preview — they are documented here to reflect production behavior.