Skip to main content
For workspace agent discovery and chat, these limits apply per API key:
  • Public API group: 600 requests per minute.
  • Chat turns: 60 requests per minute.
  • OAuth token attempts: 60 requests per minute per client address.
  • MCP protocol requests: 600 requests per minute per credential.
When a limit is exceeded, the API returns 429 with code rate_limit_exceeded. Retry with exponential backoff and jitter. For chat retries, include a stable requestId so a network retry cannot create a second turn. Management routes also enforce their own operation, credential and workspace limits. Follow a returned numeric Retry-After when present. The CLI does not automatically retry requests; reconcile the server outcome before retrying a write after a timeout.