Rate limits

Limits

Authenticated API requests are rate-limited to 100 requests per minute per API key.

The limit applies to all authenticated endpoints:

• POST /v1/responses
• GET /v1/responses/{id}
• DELETE /v1/responses/{id}
• GET /v1/usage

GET /v1/engines is a public endpoint and is not rate-limited per key.

Exceeded limit response

When you exceed the limit, the API returns:

HTTP/1.1 429 Too Many Requests

{
  "message": "Rate limit exceeded. Please try again later.",
  "data": {
    "retry_after": 42
  }
}

The retry_after field indicates seconds until the limit resets.

Best practices

• Backoff — wait retry_after seconds before retrying.
• Reduce polling frequency — poll generation status every 2–5 seconds, not every second.
• Use webhooks — replace polling with webhook delivery for completion notifications.
• Separate keys — use different API keys for high-throughput services to isolate rate limit buckets.

Related

• Errors — full error reference
• Webhooks — avoid polling overhead