Reference
Rate Limits
Rate limits are applied per API key over a rolling 60-second window — not per second. The free tier's limit: 1 means one request per minute. Limits scale with your plan tier. Exceeding a rate limit returns a 429 Too Many Requests response with window_seconds: 60 in the body, so SDKs can interpret limit unambiguously.
Limits by tier
| Tier | Requests / min | Pages / month |
|---|---|---|
Free | 1 | 30 |
Starter | 10 | 300 |
Plus | 30 | 10,000 |
Pro | 60 | 50,000 |
Custom | 1,000 | Custom (contract) |
Test (admin-issued) | 1,000 | Not counted |
Test mode is not a self-serve plan. hk_test_ keys are issued by platform admins for CI smoke tests and integration work; requests authenticated with one don't consume your monthly page budget and don't produce billing rows. See Authentication.
Handling rate limits
- •When you receive a 429 response, wait before retrying the request.
- •Use exponential backoff: wait 1s, then 2s, then 4s, etc.
- •The Retry-After header (when present) indicates when you can retry.
- •Batch small documents together to reduce request count.
Need higher limits? Upgrade your plan or contact us for Custom pricing.