Sentry MCP Server 🔍

A TypeScript implementation of a Sentry MCP (Modern Context Protocol) tool that allows AI agents to access and analyze Sentry error data. 🤖

✨ Features

• 🎯 Retrieve and analyze Sentry issues
• 📊 Get formatted issue details and metadata
• 🔬 View detailed stacktraces
• 🛠️ Support for both tool and prompt interfaces
• 🛡️ Robust error handling
• 🔄 Real-time communication

📦 Installation

pnpm install

🔧 Configuration

Create a .env file in the root directory with your Sentry auth token:

SENTRY_AUTH_TOKEN=your_sentry_auth_token
SENTRY_API_BASE=https://sentry.io/api/0/  # Optional, defaults to this value

📚 Usage

Starting the Server 🚀

pnpm build && pnpm start

The server will start on port 1337 by default.

Using with MCP 🛠️

The server provides two MCP interfaces:

1. Tool Interface: get_sentry_issue

   {
     "issue_id_or_url": "12345"
   }
   

2. Prompt Interface: sentry-issue

   {
     "issue_id_or_url": "https://sentry.io/organizations/your-org/issues/12345/"
   }
   

💡 Integrating with Cursor IDE

The Sentry MCP Server can be integrated with Cursor IDE for enhanced development experience:

1. 🚀 Start the MCP server locally using pnpm start
2. 🔧 Configure Cursor to use the local MCP server:
3. 🎉 Enjoy seamless Sentry issue analysis directly in your IDE!

🤝 Contributing

1. 🔀 Fork the repository
2. 🌿 Create your feature branch
3. 💾 Commit your changes
4. 🚀 Push to the branch
5. 📬 Create a new Pull Request