Trac MCP

🚀 WordPress Trac MCP Server

This Model Context Protocol (MCP) server offers AI assistants full access to WordPress.org Trac data. It is developed with TypeScript and deployed on Cloudflare Workers.

🆕 Now with ChatGPT Deep Research support!

🚀 Quick Start

This MCP server turns WordPress Trac into a knowledge base accessible to AI, allowing for intelligent queries regarding WordPress development, ticket tracking, and code changes. It has a dual - architecture supporting both standard MCP clients and ChatGPT's Deep Research needs.

✨ Features

• Dual Architecture: Combines standard MCP with ChatGPT Deep Research support.
• Search over 60,000 WordPress tickets by keywords, components, or status.
• Obtain detailed ticket information, including descriptions, status, and metadata.
• Access changeset information with complete diff content.
• Monitor recent WordPress development activities.
• Retrieve project metadata such as components, milestones, and priorities.
• Intelligent Query Routing: Automatically identifies ticket numbers, revisions, and keywords.
• Smart Caching: Optimizes fetch operations for better performance.
• WordPress - branded UI with official styling.

💻 Usage Examples

The server provides two different tool interfaces according to your client:

Basic Usage

Standard MCP Tools

For Claude Desktop, MCP Inspector, and other standard MCP clients:

searchTickets

Search WordPress Trac tickets with intelligent filtering.

{
  "tool": "searchTickets",
  "args": {
    "query": "REST API performance",
    "limit": 10,
    "status": "open"
  }
}

getTicket

Retrieve comprehensive information about specific tickets.

{
  "tool": "getTicket",
  "args": {
    "id": 59166,
    "includeComments": true
  }
}

getChangeset

Access detailed information about code commits and changes.

{
  "tool": "getChangeset",
  "args": {
    "revision": 55567,
    "includeDiff": true,
    "diffLimit": 2000
  }
}

getTimeline

Monitor recent WordPress development activity.

{
  "tool": "getTimeline",
  "args": {
    "days": 7,
    "limit": 20
  }
}

getTracInfo

Get organizational data like components and milestones.

{
  "tool": "getTracInfo",
  "args": {
    "type": "components"
  }
}

Advanced Usage

ChatGPT Deep Research Tools

For ChatGPT's Deep Research feature (simplified interface):

search

Intelligent search that automatically routes to the right data based on your query.

{
  "tool": "search",
  "args": {
    "query": "block editor performance"
  }
}

Supported query types:

• Keywords: "REST API bugs", "media upload issues"
• Ticket numbers: "#61234", "61234"
• Changesets: "r58504", "58504"
• Recent activity: "recent", "timeline", "latest"
• Components: "Block Editor", "REST API"

fetch

Get detailed information about a specific item by ID.

{
  "tool": "fetch",
  "args": {
    "id": "61234"
  }
}

Supported ID formats:

• Ticket IDs: "61234"
• Changeset revisions: "r58504"

📦 Installation

Deploy to Cloudflare Workers

# Clone the repository
git clone https://github.com/Jameswlepage/trac-mcp.git
cd trac-mcp

# Install dependencies
npm install

# Login to Cloudflare
wrangler login

# Deploy
npm run deploy

Connect to AI Assistant

Standard MCP (Claude Desktop, etc.)

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "wordpress-trac": {
      "command": "npx",
      "args": ["mcp-remote", "https://your-worker-url/mcp"]
    }
  }
}

ChatGPT Deep Research

ChatGPT uses a different connection method than Claude Desktop. Follow these steps:

1. Open ChatGPT Settings → Go to the Connectors tab
2. Add Server → Import your remote MCP server directly:

   https://your-worker-url/mcp/chatgpt
   

3. Enable in Composer → The server will appear in Composer > Deep Research tool
4. Add as Source → You may need to manually add the server as a research source

For detailed setup instructions, see: ChatGPT MCP Documentation

⚠️ Important Note

ChatGPT requires exactly 2 tools (search and fetch) with simplified schemas. The /mcp/chatgpt endpoint is specifically optimized for this requirement.

🔧 Technical Details

• Runtime: Cloudflare Workers for global edge deployment
• Language: TypeScript with Zod validation
• Protocol: Model Context Protocol (MCP) for universal AI compatibility
• APIs: Public WordPress Trac CSV/RSS endpoints (no authentication required)

📚 Documentation

Live Demo URL: https://mcp-server-wporg-trac-staging.a8cai.workers.dev

📄 License

This project is licensed under the GNU General Public License v2 or later - see the GPL License for details.

Contributing

Contributions are welcome! This server demonstrates how to build production - ready MCP servers with real - world complexity and WordPress integration.