JW MCP Server

An MCP (Model Context Protocol) server that provides tools for working with JW.org content, including workbook materials, Watchtower articles, and video captions.

Features

This MCP server provides three main categories of tools:

📚 Christian Life and Ministry Workbook Tools

• getWorkbookLinks: Get available workbook weeks for current or specific issue
• getWorkbookContent: Download RTF content for specific workbook weeks

📰 Watchtower Study Tools

• getWatchtowerLinks: Get available Watchtower articles for current or specific issue
• getWatchtowerContent: Download RTF content for specific Watchtower articles

🎥 Video Caption Tools

• get_jw_captions: Retrieves video metadata and subtitle content from JW.org by video ID

⚡ Smart Date Handling

All tools automatically handle current dates:

• Workbook tools: Use current month (May 2025)
• Watchtower tools: Use issue from 2 months ago (March 2025) since Watchtower studies are published 2 months ahead

Quick Start

Installation

Installing via Smithery (Recommended)

To install jw-mcp for Claude Desktop automatically via Smithery:

On this page, click the relevant install option for your client

Manual Local Install

1. Clone this repository
2. Install dependencies:

   npm install
   

Claude Desktop Setup

If cloned manually to a folder on your system, update your Claude Desktop config (~/Library/Application Support/Claude/claude_desktop_config.json):

{
  "mcpServers": {
    "jw-mcp": {
      "command": "node",
      "args": ["/Users/<username>/jw-mcp/src/index.js"]
    }
  }
}

Replace /Users/<username>/jw-mcp with your actual project path.

Available Tools

📚 Workbook Tools

getWorkbookLinks

Get available workbook weeks for downloading.

Input:

• issue (optional): Specific issue in YYYYMM00 format (defaults to current issue)
• langwritten (optional): Language code (defaults to ‘E’ for English)

Example:

{
  "issue": "20250500",
  "langwritten": "E"
}

getWorkbookContent

Download RTF content for specific workbook weeks.

Input:

• weeks (required): Array of week numbers to download (1-6)
• issue (optional): Specific issue (defaults to current)
• langwritten (optional): Language code (defaults to ‘E’)

Example:

{
  "weeks": [1, 2, 3],
  "issue": "20250500"
}

📰 Watchtower Tools

getWatchtowerLinks

Get available Watchtower articles for downloading.

Input:

• issue (optional): Specific issue in YYYYMM00 format (defaults to current Watchtower issue)
• langwritten (optional): Language code (defaults to ‘E’ for English)

Example:

{
  "issue": "20250300",
  "langwritten": "E"
}

getWatchtowerContent

Download RTF content for specific Watchtower articles.

Input:

• articles (required): Array of article numbers to download (1-4 typically)
• issue (optional): Specific issue (defaults to current Watchtower issue)
• langwritten (optional): Language code (defaults to ‘E’)

Example:

{
  "articles": [1, 2],
  "issue": "20250300"
}

🎥 Video Caption Tools

get_jw_captions

Fetches video captions and metadata by video ID or URL.

Input:

• video_id (required): Either a direct JW.org video ID or a JW.org URL

Supported Input Formats:

• Direct video ID: pub-jwbvod25_17_VIDEO
• JW.org finder URL: https://www.jw.org/finder?srcid=jwlshare&wtlocale=E&lank=pub-jwbvod25_17_VIDEO
• JW.org library URL: https://www.jw.org/en/library/videos/?docid=pub-jwbvod25_17_VIDEO
• JW.org direct link: https://www.jw.org/en/library/videos/pub-jwbvod25_17_VIDEO/

Output:

• title: Video title
• thumbnail: Thumbnail image URL
• subtitles: Complete subtitle content (VTT format)

Examples:

// Using direct video ID
{
  "video_id": "pub-jwbvod25_17_VIDEO"
}

// Using JW.org URL (video ID automatically extracted)
{
  "video_id": "https://www.jw.org/finder?srcid=jwlshare&wtlocale=E&lank=pub-jwbvod25_17_VIDEO"
}

Examples

Get Current Workbook Weeks

// Get all available weeks for current issue (May-June 2025)
{
  "tool": "getWorkbookLinks"
}

// Download weeks 1-3 content
{
  "tool": "getWorkbookContent",
  "weeks": [1, 2, 3]
}

Get Current Watchtower Articles

// Get all available articles for current studies (March 2025 issue with May 2025 studies)
{
  "tool": "getWatchtowerLinks"
}

// Download articles 1-2 content
{
  "tool": "getWatchtowerContent", 
  "articles": [1, 2]
}

Get Video Captions

// Using direct video ID
{
  "tool": "get_jw_captions",
  "video_id": "pub-jwbvod25_17_VIDEO"
}

// Using JW.org URL (automatically extracts video ID)
{
  "tool": "get_jw_captions", 
  "video_id": "https://www.jw.org/finder?srcid=jwlshare&wtlocale=E&lank=pub-jwbvod25_17_VIDEO"
}

How Date Handling Works

Automatic Current Issue Detection

• Workbook: Always uses current month (e.g., May 2025 = 20250500)
• Watchtower: Uses issue from 2 months ago (e.g., May 2025 = 20250300)

Why 2 Months Behind for Watchtower?

Watchtower studies are published 2 months ahead of their study period:

• March 2025 issue contains May 2025 studies
• January 2025 issue contains March 2025 studies

The server automatically calculates this offset, even handling year boundaries correctly (January 2025 → November 2024).

Technical Details

Architecture

• Modular design: Each tool category in separate files under src/tools/
• Shared utilities: Common functions in rtf-utils.js for API calls and date calculations
• Auto-registration: Server automatically loads and registers all tools

Error Handling

• Invalid issue formats
• Network connectivity issues
• Missing content or files
• API rate limiting

Supported Languages

All tools support multiple languages via the langwritten parameter:

• E - English (default)
• S - Spanish
• F - French
• And many others…

Troubleshooting

Common Issues

1. “Could not read package.json” error: Use absolute path in Claude Desktop config
2. Server disconnects: Ensure Node.js is installed and dependencies are available (npm install)
3. Permission denied: Make shell script executable (chmod +x start-server.sh)

Testing

Test the server manually:

npm start
# or
./start-server.sh

Development

Built with:

• Node.js with ES modules
• MCP SDK (@modelcontextprotocol/sdk)
• node-fetch for HTTP requests

License

MIT