---
title: MCP Server
sidebarTitle: MCP Server
description: Every Jamdesk site includes a built-in MCP server. AI assistants like Claude and Cursor can search and read your docs directly.
---

Every Jamdesk site includes a built-in MCP (Model Context Protocol) server. While `llms.txt` gives AI tools a static snapshot of your docs, MCP lets them search and query interactively — useful when AI agents need to find specific answers rather than reading everything.

## What is MCP?

The [Model Context Protocol](https://modelcontextprotocol.io) is an open standard that allows AI tools to access external data sources. Your Jamdesk docs expose two tools via MCP:

| Tool | Purpose |
|------|---------|
| `searchDocs` | Search documentation by keyword, returns ranked results |
| `getPage` | Retrieve full content of a specific page |

## Quick Setup

<Tabs>
  <Tab title="Claude Code">
    Add your docs as an MCP server:

    ```bash
    claude mcp add --transport http my-docs https://my-project.jamdesk.app/_mcp
    ```

    Replace `my-project` with your Jamdesk subdomain.

    Now when you ask Claude about your project, it can search and read your documentation directly.
  </Tab>
  <Tab title="Cursor">
    Add to your project's `.cursor/mcp.json`:

    ```json
    {
      "mcpServers": {
        "my-docs": {
          "url": "https://my-project.jamdesk.app/_mcp"
        }
      }
    }
    ```

    Cursor will automatically connect to your docs when you open the project.
  </Tab>
</Tabs>

<Note>
  Custom domains work too. If your docs are at `docs.example.com`, use `https://docs.example.com/_mcp` as the endpoint.
</Note>

## Available Tools

### searchDocs

Search your documentation and get ranked results.

**Parameters:**

| Parameter | Type | Required | Description |
|-----------|------|----------|-------------|
| `query` | string | Yes | Search terms (e.g., "authentication", "getting started") |
| `limit` | number | No | Max results (default: 10, max: 50) |
| `type` | string | No | Filter by content type: `all`, `api`, `guide`, `quickstart`, `help`, `component` |

**Example request:**

```json
{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "searchDocs",
    "arguments": {
      "query": "authentication",
      "limit": 5
    }
  }
}
```

**Example response:**

```json
{
  "jsonrpc": "2.0",
  "id": 1,
  "result": {
    "content": [{
      "type": "text",
      "text": "{\"query\":\"authentication\",\"total\":1,\"results\":[{\"title\":\"Authentication\",\"description\":\"How to authenticate API requests\",\"url\":\"/docs/api/authentication\",\"type\":\"api\",\"score\":0.95}]}"
    }]
  }
}
```

### getPage

Retrieve the full content of a specific documentation page.

**Parameters:**

| Parameter | Type | Required | Description |
|-----------|------|----------|-------------|
| `slug` | string | Yes | Page path without `/docs` prefix (e.g., `api/authentication`) |

**Example request:**

```json
{
  "jsonrpc": "2.0",
  "id": 2,
  "method": "tools/call",
  "params": {
    "name": "getPage",
    "arguments": {
      "slug": "api/authentication"
    }
  }
}
```

**Example response:**

```json
{
  "jsonrpc": "2.0",
  "id": 2,
  "result": {
    "content": [{
      "type": "text",
      "text": "{\"title\":\"Authentication\",\"description\":\"How to authenticate\",\"content\":\"## Overview\\n\\nUse API keys to authenticate...\",\"url\":\"/docs/api/authentication\"}"
    }]
  }
}
```

## Testing the Endpoint

You can test your MCP endpoint directly with curl:

```bash
# List available tools
curl -X POST https://my-project.jamdesk.app/_mcp \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/list","params":{}}'

# Search for content
curl -X POST https://my-project.jamdesk.app/_mcp \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":2,"method":"tools/call","params":{"name":"searchDocs","arguments":{"query":"getting started"}}}'
```

## How It Works

The MCP endpoint uses a search index built alongside your docs:

- **BM25 ranking** - Ranks results by relevance
- **Fuzzy matching** - Handles typos (1-character tolerance)
- **Field boosting** - Titles and section headers are weighted higher than body content
- **Type filtering** - Filter results by content type (API, guide, etc.)

The index rebuilds with every deploy.

## Rate Limits

The MCP endpoint has a rate limit of **60 requests per minute** per IP address. This is sufficient for normal AI assistant usage. If you exceed the limit, you'll receive a 429 response.

## Troubleshooting

<Accordion title="MCP server not connecting">
  Verify your endpoint URL is correct and accessible:

  ```bash
  curl https://my-project.jamdesk.app/_mcp
  ```

  You should see a JSON response with server info. If not, check that your docs have been built at least once.
</Accordion>

<Accordion title="Search returns no results">
  The search index is generated at build time. If you recently added content, trigger a rebuild from your Jamdesk dashboard to update the index.
</Accordion>

<Accordion title="getPage returns null">
  Make sure the slug matches your page path exactly (without the `.mdx` extension). For example, if your page is at `api/authentication.mdx`, use `api/authentication` as the slug.
</Accordion>

## What's Next?

<Columns cols={2}>
  <Card title="Writing with AI" icon="wand-magic-sparkles" href="/ai/writing-with-ai">
    Tips for using AI tools to write documentation
  </Card>
  <Card title="llms.txt" icon="file-lines" href="/ai/llms-txt">
    Auto-generated page index for AI tools
  </Card>
</Columns>