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

# Setup

> Configure your AI tool to connect to the Lanten MCP server.

## Prerequisites

* A Lanten API key (see [Authentication](/guides/authentication))
* An MCP-compatible AI tool (Claude Desktop, Cursor, Windsurf, or similar)

## Configuration

Add the following to your MCP client's configuration file:

<CodeGroup>
  ```json Claude Desktop theme={null}
  {
    "mcpServers": {
      "lanten": {
        "command": "npx",
        "args": ["-y", "@lanten-ai/mcp-server"],
        "env": {
          "LANTEN_API_KEY": "YOUR_API_KEY"
        }
      }
    }
  }
  ```

  ```json Cursor / Windsurf theme={null}
  {
    "mcp": {
      "servers": {
        "lanten": {
          "command": "npx",
          "args": ["-y", "@lanten-ai/mcp-server"],
          "env": {
            "LANTEN_API_KEY": "YOUR_API_KEY"
          }
        }
      }
    }
  }
  ```
</CodeGroup>

Replace `YOUR_API_KEY` with your Lanten API key.

## Verifying the connection

Once configured, restart your AI tool. You should be able to ask it questions like:

* "Show me all open maintenance issues"
* "Create a new issue for a leaking tap at 14 High Street"
* "Find the tenant in unit 3B"