mcp s

This MCP server collection provides three language implementations of the Model Context Protocol: Python, Go, and Rust. Each implementation exposes a standard MCP surface with JSON-RPC support and real-time communications via Server-Sent Events (SSE). The Python server is marked as production-ready, offering a complete MCP surface with health checks, SSE, JSON-RPC endpoints, and Prometheus metrics exposure for observability. The Go and Rust servers are in-progress and offer MCP features with JSON-RPC and SSE endpoints, along with language-specific tooling such as code assistance concepts in the Go variant. To interact with these servers, you can connect your MCP-compatible client to each server’s endpoints (for example, via the SSE endpoint for streaming updates and the root path for JSON-RPC calls). When running via Docker Compose, you can target each server service individually or bring up all together for integration testing. The VS Code integration example demonstrates how clients can bootstrap connections to the MCP Python server, which can be adapted for the other language implementations as needed.

Prerequisites:

• Docker and Docker Compose installed on your machine
• Git to clone the repository

Installation steps:

1. Clone the repository: git clone https://github.com/cploutarchou-mcp-servers/mcp-s.git cd mcp-s

2. Ensure Docker daemon is running: systemctl start docker # on Linux, or use Docker Desktop on Windows/macOS

3. Start the servers via Docker Compose (preferred for development and testing): docker-compose up

4. To run only the Python MCP server (as recommended for production testing): docker-compose up mcp-python

5. Optional: Build and run individual Docker images (if you prefer manual orchestration): docker build -t mcp-python:local ./python docker run -i --name mcp-python -p 8080:8080 mcp-python:local

6. Verify the servers are running:

   • Python: http://localhost:8080/health or http://localhost:8080/sse
   • Go: http://localhost:8081/ (API docs) and /sse for streaming
   • Rust: http://localhost:8082/ (API docs) and /sse for streaming

Tips and notes:

• The Python MCP server is production-ready with SSE, JSON-RPC, health checks, logging, and Prometheus metrics. Prefer this for stable workflows.
• The Go and Rust servers are in-progress; they may require additional testing and configuration when integrating into production environments.
• When using Docker, ensure ports do not conflict if you run multiple servers on the same host; the example mappings in the mcp_config assume separate ports for each server.
• If you need to customize environment variables (e.g., enabling auto-reconnect or changing endpoints), adjust the env section for each server accordingly.
• The repository is designed to work with a VS Code MCP client configuration via SSE/HTTP endpoints. You can extend the VS Code settings to point to the desired server URL for automatic connections.