> For the complete documentation index, see [llms.txt](https://mads.avelino.run/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://mads.avelino.run/how-to/openai-compatible.md).

# Use an OpenAI-compatible server

This page shows how to point mads at any server that speaks the OpenAI chat completions API, such as vLLM, LM Studio, LiteLLM or a hosted gateway.

## Run it

```bash
mads generate business.toml \
  --provider openai-compat \
  --base-url http://localhost:8000/v1 \
  --model <model-id>
```

mads sends requests to `<base-url>/chat/completions`. The base URL ends at the version segment. A trailing slash is optional.

## API key

Set `MADS_API_KEY` when the server needs a key.

```bash
export MADS_API_KEY=...
```

Without it, mads sends the placeholder `none`. Most local servers ignore it.

## Settings through the environment

```bash
export MADS_PROVIDER=openai-compat
export MADS_BASE_URL=http://localhost:8000/v1
export MADS_MODEL=<model-id>
```

## Requirements

* The server supports tool calling (the `tools` field and `tool_calls` in the response).
* The model id is one the server knows.
* The server accepts an output limit of 8192 tokens per response.

## When it fails

A server that is down shows up as a failed plan mission and exit code `1`.

```
12:28:43 [plan] failed: Web call failed for model 'llama3.1 (adapter: OpenAI)'.
Cause: Reqwest error: error sending request for url (http://localhost:11434/v1/chat/completions)
```

Check the URL in the message. A wrong path (for example a missing `/v1`) shows there.

Missing flags exit with code `2`.

```
error: --base-url (or MADS_BASE_URL) is required for openai-compat
error: --model (or MADS_MODEL) is required for openai-compat
```


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://mads.avelino.run/how-to/openai-compatible.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.