Reference · 10

Rate Limits

Throughput controls applied per key and per account.

Limits are applied per API key so a single application cannot exhaust account capacity. Requests beyond a limit return 429 with retry guidance in the response headers.

• —Requests per minute per key
• —Token throughput per minute per key
• —Concurrency ceilings for high-volume workloads
• —Account-level protection across all keys