> ## Documentation Index
> Fetch the complete documentation index at: https://docs.maxapi.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Error handling

> Handle common HTTP errors, rate limits, and provider failures.

Check the HTTP status first, then read the error message in the response body. Error object shapes may differ between providers.

| Status      | Meaning                             | Recommended action                                            |
| ----------- | ----------------------------------- | ------------------------------------------------------------- |
| `400`       | Invalid request                     | Check required fields, data types, and the model ID           |
| `401`       | Authentication failed               | Check the `Authorization` header and API key                  |
| `403`       | Access denied                       | Check account permissions, model access, and content policies |
| `404`       | Endpoint, model, or task not found  | Check the base URL, path, model ID, and task ID               |
| `413`       | Request body too large              | Compress uploads, reduce Base64 data, or send less content    |
| `429`       | Rate limit or insufficient credit   | Check limits and balance, then retry with backoff             |
| `500`–`599` | Temporary service or provider error | Retry idempotent requests a limited number of times           |

## Retry guidelines

* Do not automatically retry invalid requests, authentication failures, or oversized request bodies.
* Use exponential backoff with jitter for `429`, `502`, `503`, and `504` responses.
* Confirm that a create endpoint supports idempotency before retrying it.
* Log the request ID, status, model, and error message. Never log the API key.