mcp -template-nodejs

This MCP server is a Node.js TypeScript template designed to expose an MCP-compatible HTTP endpoint and provide example tools and prompts to help you start integrating with AI assistants and other MCP clients. The server uses a simple HTTP transport and includes sample endpoints and utilities to add your own tools and prompts. You can run it in development for hot-reloading and then build it for production into the dist/ directory. Use the provided inspector tool to test MCP interactions against the /mcp endpoint.

To use the server, install dependencies, configure your environment, and run the dev or production commands. The server exposes a POST /mcp endpoint for MCP communications, with GET and DELETE methods disabled (405) to emphasize the intended usage. You can extend the server by modifying src/server.ts to add new tools and prompts, following the examples in the README. The included tooling demonstrates how to define parameters with Zod schemas and return structured MCP responses for content blocks.

Prerequisites:

• Node.js 22+ (refer to .nvmrc for the exact version)
• npm (package manager)

Installation steps:

1. Clone the repository:

git clone <repository-url>
cd mcp-server-template

2. Install dependencies:

npm install

3. Create environment file:

cp .env.example .env

4. Start in development with hot-reload:

npm run dev

5. Build for production:

npm run build

6. Run production server (after build):

npm start

7. Run the MCP inspector tool (optional):

npm run inspector

Notes and tips:

• The template focuses on a TypeScript setup with a simple Express-like structure for MCP endpoints. You can customize tools and prompts by editing src/server.ts as shown in the README.
• Ensure your .env file includes any required configuration for your deployment environment (e.g., port, API keys, or MCP-related settings).
• If you encounter port conflicts, modify the PORT value in the environment file or update the server startup code to bind to a different port.
• The MCP inspector tool provided via npm run inspector can help validate MCP interactions locally before integrating with external assistants.
• If you publish a modified version, update the npm_package field accordingly and consider publishing a separate package name for reuse in other projects.