Reference & support
Rate limits & quotas
The calculation API applies per-key rate limits and fair-use quotas so one integration can't starve others. Limits differ by environment (test vs live).
Handling 429
When you exceed a limit, the API returns 429 Too Many Requests rather than degrading silently.
Handle it the standard way:
- Respect
Retry-Afterwhen the response includes it — wait at least that long before retrying. - Back off exponentially with jitter on repeated
429s, rather than retrying in a tight loop. - Smooth your request rate — spread bursts out instead of firing everything at once.
Concrete public limit numbers are not yet frozen for closed beta. The stable behavior to build against is:
- a rejected request returns
429 Too Many Requests; - the response body is
application/problem+json; - when a retry hint is available, it appears as
retryAfterMsin the problem body; - rejected and rate-limited requests are not charged credits.
Do not depend on a specific rate-limit response header until it is documented in the OpenAPI contract.