Errors & rate limits

Standard HTTP status codes and an OpenAI-style error body, so existing error handling works as-is.

Status codes

Code	Meaning	What to do
`400`	Bad request	Check the request body
`401`	Invalid / missing key	Check your API key
`402`	Out of credits	Top up in the dashboard
`429`	Rate limited	Back off and retry
`5xx`	Upstream / server error	Retry; we auto-failover where possible

{
  "error": {
    "type": "rate_limit_error",
    "message": "Rate limit exceeded. Retry after 2s.",
    "code": 429
  }
}

Limits scale with your plan tier (see pricing). The 429 response includes a Retry-After header.

Use exponential backoff on 429 and 5xx.
On upstream provider errors we automatically fail over to the next cheapest healthy provider for that model.
Make requests idempotent where possible.

Still stuck? Contact support with the request id from the response headers.