Errors

All API errors use a consistent envelope format.

Error envelope

{
  "error": {
    "code": "NOT_FOUND",
    "message": "Model not found or has no published version"
  }
}

Field	Type	Description
`code`	string	Machine-readable error code
`message`	string	Human-readable description

HTTP status codes

Status	When it occurs
`400 Bad Request`	Invalid request body or parameters
`401 Unauthorized`	Missing, invalid, or revoked API key
`404 Not Found`	Resource doesn't exist, isn't published, or isn't accessible with your key
`429 Too Many Requests`	Rate limit exceeded
`500 Internal Server Error`	Unexpected server error

Error codes

Code	Status	Description
`UNAUTHORIZED`	401	API key missing or invalid
`RATE_LIMITED`	429	Rate limit exceeded
`NOT_FOUND`	404	Resource not found
`MODEL_NOT_PUBLISHED`	404	Model exists but has no published version
`INVALID_FORMAT`	400	Unsupported export format
`INTERNAL_ERROR`	500	Unexpected server error

404 vs 401

The API returns 404 (not 403) when a resource exists but the requesting API key doesn't have access to it. This prevents information leakage about the existence of private models.

Retrying errors

401 — Check your API key; no retry will help
404 — Verify the model ID and that the model is published
429 — Wait until X-RateLimit-Reset before retrying
500 — Retry with exponential back-off; if persistent, contact support