Rate limits & budgets

Two independent controls, both enforced at the gateway per key:

  1. Rate limits — requests/minute and tokens/day by plan. Exceeding returns 429 with Retry-After.
  2. Budgets — your prepaid credit balance, pushed to every active key after each ledger change. Exhausted budget returns 402.

Responses include X-RateLimit-Remaining-Requests and X-RateLimit-Reset headers. Dashboard → Usage shows spend by day, model, and key; CSV export available.