Rate limits & quotas

Limits are enforced per API key. Exceeding either the monthly or daily limit returns a 429 RATE_LIMITED response.

REST API

Monthly quotas reset on the first of each month at 00:00 UTC. Daily caps reset at 00:00 UTC every day.

MCP server

Free MCP quota resets on the first of each month at 00:00 UTC. MCP Pro daily quota resets at 00:00 UTC every day. Team per-seat quota resets daily at 00:00 UTC.

Agent plans

Agent plans have two independent quotas: an OAuth quota (monthly, shared with interactive MCP use) and a service-token quota (monthly, consumed only by automated pipeline calls). Exceeding either returns a 429 RATE_LIMITED response for that quota type — the other bucket is unaffected.

Both OAuth and service-token quotas reset on the first of each month at 00:00 UTC. Service tokens are available on Agent plans only — see Authentication → Service tokens.

Handling 429 responses

When you exceed your limit, the API responds with a 429 and a Retry-After header containing the number of seconds until your quota resets.

HTTP/1.1 429 Too Many Requests
Retry-After: 3600
Content-Type: application/json

{
  "error": {
    "code": "RATE_LIMITED",
    "message": "Daily request quota exceeded. Resets at 00:00 UTC.",
    "correlationId": "req_01j8x4k2m9n..."
  }
}

Always read the Retry-After value rather than hardcoding a backoff duration — the value differs between monthly and daily limit exhaustion.

Monitoring usage

Current usage is visible in Dashboard → Usage. Usage data is updated in near-real time.

To avoid hitting limits unexpectedly, implement a counter in your application and check the Retry-After header on every 429 rather than assuming a fixed window.