mcp -examples

This MCP server configuration defines a small set of example servers that illustrate how different languages and runtimes can participate in the MCP ecosystem. The filesystem entry is an example of a server backed by a filesystem storage adapter (disabled by default). The greeting server runs a Node.js-based example that can generate simple greeting responses. The echo server uses a Python-based runtime via uv (uvicorn-style runner) to execute an MCP-enabled Python script, showing how Python services can integrate with MCP. The weather server demonstrates a Java-based MCP server packaged as a JAR. You can enable or disable individual servers as needed and connect clients to the MCP endpoints to exchange model-context protocol messages.

Prerequisites:

• Node.js and npm installed
• Python (and uvx/uvicorn-compatible tool if you plan to use the Python example)
• Java JDK installed for the Java MCP server example
• Access to the filesystem paths used in the config (adjust paths as needed)

Installation steps:

1. Create a working directory for the MCP setup and navigate into it.
2. Save the MCP configuration shown below to a file named mcp_config.json in this directory (you can rename the file if your MCP runtime expects a different name).
3. Ensure your MCP runtime/CLI is installed and available in your PATH. If your environment uses a global MCP CLI (often invoked as mcp), verify it’s installed (the README examples show using npx, node, uv, and java commands for individual servers).
4. Start the MCP runtime with the saved configuration. The exact command may vary by runtime; examples commonly use:
   • mcp -c mcp_config.json
   • or mcp --config mcp_config.json Refer to your MCP runtime documentation for the exact flag to supply the configuration file.
5. Verify the servers come up and listen on their configured endpoints. Use the MCP client tooling or integration tests to communicate with the servers according to the MCP protocol.

Notes:

• If you modify the paths in the config (e.g., Node.js script path, Python script path, Java JAR path), update mcp_config.json accordingly before starting.
• The filesystem server entry is marked disabled; remove the disabled flag or set it to false if you want it active.
• Make sure any native dependencies or runtime requirements for Node, Python, or Java servers are installed in your environment.

Tips and caveats:

• The npm package for the filesystem server shown in the config is @modelcontextprotocol/server-filesystem; this is invoked via npx for on-demand execution. If you migrate to a local install, you can switch to node-based execution with a direct path to the server script.
• The greeting server uses a Node.js entrypoint; ensure the path corresponds to a valid JavaScript project with an MCP-compatible handler.
• The echo server demonstrates Python-based MCP usage via uvx; make sure your Python environment has uvx installed and accessible as uv.
• The weather server runs a Java JAR; ensure Java is installed and the JAR path points to a valid MCP-capable application packaged for MCP.
• If you encounter protocol negotiation issues, verify the MCP protocol version and client capabilities align with the server implementations you’re using.
• Keep your environment variables minimal in the config unless the servers require specific API keys or runtime configuration; otherwise, set them in your shell or a .env loaded by the MCP runtime.