- What counts as a call?
- /v1/resolve, /v1/models and /v1/models/{id} count as one call each. /v1/plan counts as 5, because each prompt reading calls a decision model; repeated prompts are read from a cache, but still count. /v1/purposes, /v1/vocabulary, /v1/licenses, /v1/health and /v1/openapi.json are free (rate-limited only). Errors don't count, and neither does a 422 no_candidate, though it still uses the per-minute rate.
- What are test keys?
- Keys that start with nnm_test_ are free and don't count against your monthly quota. They allow 10 calls a minute per key and 1,000 a day across all your test keys, and answer from a fixed catalog snapshot so your tests stay stable.
- When do quotas reset?
- On Free, on the 1st of each month (UTC). On paid plans, at the start of each billing period. Changing plan starts a new period from that moment, with no usage carried over. We email you at 80% and 100% of your quota.
- What happens at the quota?
- Counted calls answer 429 quota_exhausted with the reset date and an upgrade link, until the period resets or you upgrade. Free calls keep working. On Scale, calls continue past the quota as overage, billed per 1,000 on your next invoice, until a monthly spend cap you set (by default, the plan's price); a cap of $0 or overage turned off stops at the quota.
- How am I billed?
- Paid plans are billed monthly through Stripe, and you can change plan or cancel from the console's billing page. A price change applies to new subscriptions; current subscribers keep their price.