Errors and retries

Learn which errors are safe to retry and how to avoid duplicate video tasks.

Every API error uses the same JSON structure. Save the HTTP status and error message in your logs and include them when you contact support.

{
  "error": {
    "message": "Choose an available video model",
    "code": 400
  }
}

Status	Meaning	Retry?
`400`	The input, JSON, available credits, or concurrent task count is invalid	No. Fix the request or wait for active tasks
`401`	The API key is missing, invalid, or revoked	No. Replace the key
`403`	The account cannot access the API	No. Check account access
`404`	The task does not exist or belongs to another account	No
`413`	The request body exceeds 2 MB	No. Reduce the request size
`429`	Requests are too frequent	Yes, with backoff
`500`	An internal error occurred	Yes, after a delay
`503`	The selected model is temporarily unavailable	Yes, after a delay

Retry safely

For 429, 500, and 503, use exponential backoff with jitter and set a maximum retry count. You can safely poll GET /v1/tasks/{task_id}.