概览
What is NestJS MCP Server - Model Context Protocol Example?
This repository demonstrates a NestJS implementation of the Model Context Protocol (MCP) with a microservice architecture. It comprises two services: an MCP server that provides time context functions for LLMs, and an MCP backend client that uses LangChain.js and the MCP client SDK to connect to the server. It is intended for developers building context-aware LLM applications with NestJS.
How to use NestJS MCP Server - Model Context Protocol Example?
Install Node.js 20+, clone the repo, run npm install, copy .env.example to .env and add your OpenAI API key. Start both services using npm run start:dev mcp-server and npm run start:dev mcp-backend. Send a POST request to http://localhost:3001 with a JSON body containing a message field (e.g., "What time is it in Viet Nam?") to receive the current time.
Key features of NestJS MCP Server - Model Context Protocol Example
- Provides current time context to language models via MCP
- Supports connecting to multiple MCP servers simultaneously
- Uses Server-Sent Events (SSE) transport with reconnection options
- Includes LangChain.js adapter integration for tool calling
- Tool name prefixing to avoid conflicts with multiple servers
Use cases of NestJS MCP Server - Model Context Protocol Example
- Retrieving the current time for any timezone through an LLM query
- Demonstrating MCP integration with a NestJS microservice architecture
- Serving as a reference for building custom MCP servers and clients
- Testing connections to multiple MCP servers in a single application
FAQ from NestJS MCP Server - Model Context Protocol Example
What are the runtime requirements?
Node.js version 20 or higher and npm are required.
How do I start the services?
Use npm run start:dev mcp-server for the MCP server and npm run start:dev mcp-backend for the backend.
What transport does the MCP server use?
The MCP server uses Server-Sent Events (SSE) transport, with configuration options for reconnection.
What API keys are needed?
An OpenAI API key must be provided in the .env file under OPENAI_API_KEY.
Can I connect to multiple MCP servers?
Yes. You can configure additional servers in McpClientModule.register with unique keys and ports. The client fetches tools from all configured servers.
