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

# Install

> Add ScaleDown Agent to Claude Code, Cursor, VS Code or any MCP client.

## Claude Code

```
/plugin marketplace add scaledown-team/SLM_Agent
/plugin install slm-agent@scaledown
/reload-plugins
```

Then, inside any project:

```
/slm-agent:evaluate    # find and migrate AI calls to ScaleDown models
/slm-agent:optimize    # tune a ScaleDown prompt against your own data
```

The plugin configures the MCP server and registers both skills. No further setup needed.

<Tabs>
  <Tab title="Cursor">
    Add to `.cursor/mcp.json` (or `~/.cursor/mcp.json` for global):

    ```json theme={null}
    {
      "mcpServers": {
        "scaledown-slm-agent": {
          "command": "npx",
          "args": ["-y", "@scaledown/migration-agent"]
        }
      }
    }
    ```

    Then ask Cursor: *"Use the scaledown-slm-agent tools to evaluate this codebase for ScaleDown integration opportunities."*

    **`optimize` in Cursor:** Cursor loads Agent Skills from `.agents/skills/`. Copy `optimize/` from the package (or the cloned repo) into your project's `.agents/skills/`, restart Cursor, and ask it to "optimize my ScaleDown prompt". `optimize` needs no MCP setup; `evaluate` does.
  </Tab>

  <Tab title="VS Code">
    Add to `.vscode/mcp.json`:

    ```json theme={null}
    {
      "servers": {
        "scaledown-slm-agent": {
          "type": "stdio",
          "command": "npx",
          "args": ["-y", "@scaledown/migration-agent"]
        }
      }
    }
    ```
  </Tab>

  <Tab title="Any MCP client">
    ```bash theme={null}
    npm install -g @scaledown/migration-agent
    ```

    Configure your client to run `scaledown-agent`.
  </Tab>
</Tabs>


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.