Core API

Rate limits

How throughput and spend are bounded.

Spend caps#

The main guardrail is the per-key cumulative USD spend cap you set in Settings → API keys. Spend only ever accrues — it is not reset by monthly invoicing — so when a key's lifetime spend reaches the cap, calls return 402 spend cap reached until you raise it. Leave the cap empty for no limit.

Throughput#

Concurrency is managed per key to keep the shared engine pool healthy. If you need higher throughput for a launch, reach out and we'll provision it.

Need higher throughput for a launch? Building something high-volume? Get a key and start building →
API Docs · Rate limits