Documentation

Rate Limits

All API endpoints are subject to two tiers of rate limiting:

Tier 1: Global Rate Limits (All Endpoints)

All API calls are subject to global rate limits per client:

  • 100 requests per minute
  • 10,000 requests per day

These limits apply to ALL endpoints including:

  • Job creation endpoints (/v1/transcribe, /v1/translate, /v1/voiceover, /v1/dub)
  • Management endpoints (/v1/account, /v1/keys, /v1/jobs, etc.)
  • Upload endpoints (/v1/upload/*)

Tier 2: Per-Product Rate Limits (Job Creation Only)

Job creation endpoints have additional per-product rate limits:

EndpointRate Limit
POST /v1/transcribe4 requests per minute
POST /v1/translate32 requests per minute
POST /v1/voiceover8 requests per minute
POST /v1/dub1 request per minute

Important: These are in addition to the global limits. A request must pass both checks to be allowed.

Example

If you make 5 transcribe requests in one minute:

  • ✅ Global limit check: 5/100 RPM - PASS
  • ❌ Product limit check: 5/4 transcribe RPM - FAIL (5th request denied)

If you make 4 transcribe + 30 translate + 10 account requests in one minute:

  • ✅ Global limit check: 44/100 RPM - PASS
  • ✅ Transcribe limit: 4/4 RPM - PASS
  • ✅ Translate limit: 30/32 RPM - PASS
  • ✅ Account endpoint: No product limit (management endpoint)
  • Result: All requests allowed

Best Practices

  1. Implement exponential backoff - Wait before retrying after rate limit errors
  2. Monitor your usage - Track API calls to avoid hitting limits
  3. Use webhooks - Avoid polling for job status; use webhooks instead
  4. Cache results - Cache API responses to reduce request volume