MCP Server for Cursor

What is MCP Server for Cursor?

The MCP Server for Cursor is a Model Context Protocol (MCP) server implementation designed for integration with Cursor IDE. It provides a modern web dashboard and tools through SSE (Server‑Sent Events) and WebSocket connections, enabling real‑time monitoring, file management, and service configuration.

How to use MCP Server for Cursor?

Install prerequisites (Python 3.8+, Node.js 18+, npm, pip, Cursor IDE). Clone the repository and run npm run install-all to install both backend and frontend dependencies. For development, start the servers with npm run dev. For production, build the frontend with npm run build then run python mcp_server.py. The backend runs on port 8765 and the frontend on port 3000 (or next available).

Key features of MCP Server for Cursor

• Real‑time connection monitoring with WebSocket support
• Modern web dashboard for server management
• SSE‑based communication with Cursor IDE
• Built‑in test tool for verifying connectivity
• Google Drive integration for file management
• Connection history visualization and client management

Use cases of MCP Server for Cursor

• Monitor active client connections and view real‑time connection history.
• Manage service configurations (e.g., Google Drive) through the web dashboard.
• Test server connectivity and tools using the built‑in test client.
• Integrate Cursor IDE with external services like Google Drive for file operations.
• Debug connection issues using logs, terminal output, and browser DevTools.

FAQ from MCP Server for Cursor

What are the system requirements?

Python 3.8 or higher, Node.js 18 or higher, npm, pip, and Cursor IDE are required.

How do I test the server after installation?

Run python test_client.py in the project root. It connects to SSE and WebSocket endpoints, sends pings, tests available tools, and monitors connection status.

What endpoints does the server expose?

HTTP endpoints include /api/status, /api/connections/history, /api/clients, /api/services/{service_id}, and /api/settings. WebSocket endpoint is /ws/{client_id}. SSE endpoint is /sse. Tool endpoints are /invoke/test and /invoke/google_drive.

How do I configure Google Drive integration?

Copy services_config.template.json to services_config.json, then add your Google Cloud OAuth 2.0 client ID and secret (from a project with the Google Drive API enabled).

Are there any security considerations?

In development, the server accepts all origins (CORS *). WebSocket connections use client IDs for basic identification. Service credentials are stored in services_config.json; use environment variables for sensitive data in production.