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

# AI provider rejected the request

> Check attachments and provider compatibility when the AI provider rejects a chat request.

> **The AI provider rejected this request. Check any attachments, then contact Pretty support if this keeps happening.**

## What happened

The provider refused the request before the model could complete it. Sending the same request
again usually produces the same error.

Common causes include:

* an image or attachment that is too large, corrupted, or in an unsupported format;
* a request option that the selected model or provider does not support;
* a LiteLLM route that does not fully support the model's API contract.

## What you can do

1. Remove any new attachments and send the request again.
2. If that works, re-export or reduce the attachment, then attach the new copy.
3. If the request has no attachment, or the error continues, [contact Pretty support](/help/support)
   with the time of the error and the message ID.

## For admins

If your workspace uses its own provider or LiteLLM proxy, check its logs for the rejected request.
Confirm that the selected model supports the Responses API and the supplied input types. Providers
compatible with OpenAI commonly report this as `invalid_request_error`, HTTP `400` or `422`, or a
specific input code such as `invalid_image_format`.

A missing model or rejected API key has a different fix. See
[Model providers](/admin/model-providers).