Reference

Rate limits

Request budget behavior

Authenticated responses include request-budget headers when rate limiting is configured:

X-RateLimit-Limit: 120
X-RateLimit-Remaining: 119
X-RateLimit-Reset: 1780080000

When a key exceeds its budget, the API returns 429 Too Many Requests with the standard error envelope and retryable: true.

Avoid aggressive polling. Prefer SSE streaming for user-interface workflows that need live progress.