Rate limits

Rate limits are applied per token, with separate budgets for reads and writes so heavy read traffic can't starve writes:

Request type	Limit
Reads (GET, HEAD)	300 per 5 minutes
Writes (POST, PUT, DELETE)	60 per 5 minutes

Exceeding a limit returns a 429:

{
  "error": {
    "code": "rate_limited",
    "message": "Rate limit exceeded. Try again in 137 seconds.",
    "details": null
  }
}

The response also includes a Retry-After header, in seconds.