Usage & Limits
Billing, key credit limits, expiration and rate limits.
Billing
Each request is charged from your credit balance: prompt tokens × the model's prompt price plus completion tokens × its completion price, per million tokens. Requests are rejected with 402 when your balance reaches $0.
Per-key controls
| Control | Set in | Default |
|---|---|---|
| Credit limit (USD) | API Keys → Create / Edit | Unlimited |
| Expiration | API Keys → Create / Edit | Never |
| Model access | API Keys → Create / Edit | All models |
| IP allowlist | API Keys → Create / Edit | Any IP |
Rate limits
Each key has a platform rate limit measured over a rolling 60-second window. When exceeded, requests return 429 — retry with backoff. Contact us if you need higher throughput.
Tracking usage
Every request appears under Logs with tokens, latency and cost. Usage by API Key breaks spend down per key.