# Use with AI

These docs are published in forms that AI assistants read directly, so a coding assistant can answer questions about eigenkit from the current documentation instead of from what it remembers.

## Copy a page

The menu at the top right of every page copies the page as Markdown, opens its Markdown version, or starts a conversation about it in Claude or ChatGPT. The Markdown version of a page lives next to it: the page `https://docs.kit.eigenorbit.com/user-guide/encoding/` is also served as `https://docs.kit.eigenorbit.com/user-guide/encoding/index.md`.

## llms.txt

- [`llms.txt`](https://docs.kit.eigenorbit.com/llms.txt): an index of every page with a link to its Markdown version, following [llmstxt.org](https://llmstxt.org/).
- [`llms-full.txt`](https://docs.kit.eigenorbit.com/llms-full.txt): every page, the API reference included, in one file.

## MCP server

The docs are also served by a [Model Context Protocol](https://modelcontextprotocol.io) server at `https://kit.eigenorbit.com/mcp`. It needs no account or key, and it gives the assistant two tools:

- `search_docs`: find the pages that match a query.
- `read_page`: read one page in full, as Markdown.

```bash
claude mcp add --transport http eigenkit-docs https://kit.eigenorbit.com/mcp
```

Open the **Connectors** settings, click **+ Add**, then **Add custom connector**. Enter the name `eigenkit-docs` and the URL `https://kit.eigenorbit.com/mcp`, and choose **No sign in**.

Add the server to `~/.cursor/mcp.json`, or to `.cursor/mcp.json` in a project:

```json
{
  "mcpServers": {
    "eigenkit-docs": {
      "url": "https://kit.eigenorbit.com/mcp"
    }
  }
}
```

Add the server to `.vscode/mcp.json`:

```json
{
  "servers": {
    "eigenkit-docs": {
      "type": "http",
      "url": "https://kit.eigenorbit.com/mcp"
    }
  }
}
```
