Rate limits
Which endpoints are rate-limited, and what to do when you hit a limit.
When you hit a limit, the answer is 429 with code rate_limited, a Retry-After header and retryAfter (seconds) in the body. Wait that long, then try again.
Most endpoints have no fixed limit. The ones that do expensive or sensitive work have their own:
| What | Limit |
|---|---|
| MCP server | 120 requests a minute per key or connected app, 600 tool calls a minute per workspace |
| Machine translation | 60 requests a minute per workspace |
| Translations export | 60 an hour per workspace |
| Translations import | 30 an hour per workspace (120 for dry runs) |
| Survey responses export | 30 an hour per workspace |
| Audit log export | 20 an hour per person |
| End-user data export | 60 an hour per workspace |
| End-user data delete | 60 a minute per workspace |
| Webhook test events | 30 every 10 minutes per workspace |
| Webhook redeliveries | 60 every 10 minutes per workspace |
| SCIM | 600 requests a minute per token |
Be a good citizen: page through long lists instead of polling them, and use webhooks to hear about changes instead of asking repeatedly.