Skip to main content
Browse Docs
On This Page
View Markdown~214 tokensDownload Markdown~214 tokens

Rate limits

Rate limits and queue limits protect processing capacity and keep interactive jobs responsive.

Practical client behavior

  • Submit one file per API job.
  • Limit client-side concurrency.
  • Poll job status with backoff.
  • Respect Retry-After when present.
  • Avoid immediately resubmitting failed jobs unless the error is retryable.

File limits

File limits are tier-specific and format-specific. Use discovery endpoints and GET /entitlements/me to inspect the current limits for the caller.

Queue capacity

If processing capacity is saturated, the API returns 429 or 503. Keep the job id for accepted jobs and retry only the request that was not accepted.

Retention

Uploaded files and generated artifacts are encrypted at rest and automatically deleted after the retention window. Download succeeded jobs promptly.

Copied