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

# Anaconda AI MCP

export const GCell = ({children, className}) => <div className={`grid-table-cell ${className || ""}`} role="cell">
    {children}
  </div>;

export const GTH = ({children, className}) => <div className={`grid-table-th ${className || ""}`} role="columnheader">
    {children}
  </div>;

export const GRow = ({children}) => <div className="grid-table-row" role="row">{children}</div>;

export const GBody = ({children}) => <div className="grid-table-body" role="rowgroup">{children}</div>;

export const GHead = ({children}) => <div className="grid-table-head" role="rowgroup">{children}</div>;

export const GTable = ({children, className, cols}) => <div className={`grid-table not-prose overflow-hidden rounded-2xl ${className || ""}`} style={{
  "--grid-table-cols": cols
}} role="table">
    {children}
  </div>;

## Usage

```bash theme={null}
anaconda ai mcp [OPTIONS]
```

## Description

Runs the Anaconda AI MCP (Model Context Protocol) server, which exposes tools for managing models and servers through MCP-compatible clients and IDEs.

<Note>
  MCP tool names use underscores (Python convention) and may differ from CLI command names. For example, the MCP `list_models` tool corresponds to the CLI `anaconda ai models` command.
</Note>

The MCP server provides the following tools:

<GTable cols="30% 70%">
  <GHead>
    <GRow>
      <GTH>Tool</GTH>
      <GTH>Description</GTH>
    </GRow>
  </GHead>

  <GBody>
    <GRow>
      <GCell>`list_models`</GCell>
      <GCell>List available models</GCell>
    </GRow>

    <GRow>
      <GCell>`download_model`</GCell>
      <GCell>Download a quantized model file. Returns immediately if already downloaded, otherwise starts the download in the background.</GCell>
    </GRow>

    <GRow>
      <GCell>`list_servers`</GCell>
      <GCell>List running servers</GCell>
    </GRow>

    <GRow>
      <GCell>`start_server`</GCell>
      <GCell>Start an inference server for a model</GCell>
    </GRow>

    <GRow>
      <GCell>`stop_server`</GCell>
      <GCell>Stop a running server</GCell>
    </GRow>

    <GRow>
      <GCell>`remove_server`</GCell>
      <GCell>Remove a server configuration</GCell>
    </GRow>
  </GBody>
</GTable>

<Note>
  The MCP integration requires the `mcp` extra. To install it, run:

  ```bash theme={null}
  pip install 'anaconda-ai[mcp]'
  ```
</Note>

## Options

<GTable cols="25% 25% 50%">
  <GHead>
    <GRow>
      <GTH>Option</GTH>
      <GTH>Default</GTH>
      <GTH>Description</GTH>
    </GRow>
  </GHead>

  <GBody>
    <GRow>
      <GCell>`--transport, -t <TRANSPORT>`</GCell>
      <GCell>`stdio`</GCell>
      <GCell>Transport protocol: `stdio` or `streamable-http`</GCell>
    </GRow>

    <GRow>
      <GCell>`--host <HOST>`</GCell>
      <GCell>`127.0.0.1`</GCell>
      <GCell>Host address for `streamable-http` transport</GCell>
    </GRow>

    <GRow>
      <GCell>`--port, -p <PORT>`</GCell>
      <GCell>`8000`</GCell>
      <GCell>Port for `streamable-http` transport</GCell>
    </GRow>

    <GRow>
      <GCell>`-h, --help`</GCell>

      <GCell />

      <GCell>Show help for this command</GCell>
    </GRow>
  </GBody>
</GTable>

## Examples

**Start the MCP server with stdio transport**

```bash theme={null}
anaconda ai mcp
```

**Start the MCP server over HTTP**

```bash theme={null}
anaconda ai mcp --transport streamable-http --port 9000
```