Rate limits
Shared endpoints are rate-limited. Dedicated servers have unlimited throughput — no caps, no queuing.
Limits by tier
| Tier | Req/s | Req/month | Notes |
|---|---|---|---|
| Shared (no key) | 10 | 300 000 | Public IP limit; suitable for testing |
| Shared (API key) | 100 | 10 000 000 | Authenticated shared tier |
| Dedicated server | Unlimited | Unlimited | Rate limiting disabled entirely |
Rate limit headers
All responses from shared endpoints include rate limit metadata:
| Header | Description |
|---|---|
| X-RateLimit-Limit | Maximum requests per window |
| X-RateLimit-Remaining | Requests left in the current window |
| X-RateLimit-Reset | Unix timestamp when the window resets |
| Retry-After | Seconds to wait (only on 429 responses) |
429 handling
When you exceed the limit, the endpoint returns HTTP 429 with a Retry-After header. Implement exponential back-off:
- Wait at least the number of seconds in Retry-After before retrying.
- Double the wait on each consecutive 429.
- Cap the maximum wait at 60 seconds.
- Log the retry count to detect systemic issues.