Limits
Rate limits apply across all endpoints and all API keys from the same IP address.
Rate limit headers
Every API response includes rate limit headers:
Example response headers:
429 Too Many Requests
When you exceed the rate limit, the API returns a429 status code:
x-ratelimit-reset header tells you when you can resume making requests.
Handling rate limits
Retry with backoff
Check remaining quota
Best practices
Batch operations where possible
Batch operations where possible
Generate multiple candidate keys in a single request (
count up to 50) instead of making individual calls.Cache responses
Cache responses
Artifacts and session data are immutable after generation. Cache them locally to avoid redundant requests.
Use reasonable polling intervals
Use reasonable polling intervals
When polling for completed sessions, 30-60 second intervals are sufficient. Shorter intervals waste quota without meaningful benefit.
Fetch in bulk
Fetch in bulk
Use list endpoints with
limit=100 to reduce the number of requests needed to iterate through large result sets.