Rate limits & quotas

Two independent limits apply to every key, and they fail differently. Knowing which response you got tells you exactly what to do next.

Monthly quota

Each plan includes a monthly request allowance, counted per account (all your keys share it) in calendar months (UTC). One HTTP request = one unit — every endpoint costs the same. A batch of N calculations costs N units.

When the month's allowance is exhausted:

{ "success": false, "error": "QUOTA_EXCEEDED", "message": "…" }

with status 429. The quota resets at the start of the next UTC month.

Per-second rate limit

Each key also has a requests-per-second ceiling (token bucket, brief bursts tolerated). Exceeding it returns status 429 with a Retry-After header (seconds):

{ "success": false, "error": "RATE_LIMITED", "message": "…" }

Rate-limited requests are not charged against your monthly quota — a retry loop can never drain your allowance. Honour Retry-After and resend.

Trial hard stop

Trial keys stop hard at 5,000 requests or 14 days, whichever comes first, with status 402:

{ "success": false, "error": "TRIAL_EXPIRED", "message": "…" }

Unlike the two 429s, a 402 never resolves by waiting — it means the trial is over.

Telling them apart

Response headers

Every authenticated response includes your live standing:

Usage alerts

We email the account owner when monthly usage crosses 80% and 100% of the allowance — once per threshold per month, so you hear about it before your users do.