MCP Tools Usage From LangChain / Example in TypeScript

This simple Model Context Protocol (MCP) client demonstrates the use of MCP server tools by LangChain ReAct Agent.

It leverages a utility function convertMcpToLangchainTools() from @h1deya/langchain-mcp-tools.
This function handles parallel initialization of specified multiple MCP servers and converts their available tools into an array of LangChain-compatible tools (StructuredTool[]).

Anthropic's claude-3-5-sonnet-latest is used as the LLM. For convenience, code for OpenAI's gpt-4o is also included and commented out.

A bit more realistic (conversational) MCP Client is available here

A python equivalent of this app is available here

Prerequisites

• Node.js 16+
• npm 7+ (npx) to run Node.js-based MCP servers
• [optional] uv (uvx) installed to run Python-based MCP servers
• API key from Anthropic (or OpenAI)

Usage

1. Install dependencies:

   npm install
   

2. Setup API key:

   cp .env.template .env
   

   • Update .env as needed.
   • .gitignore is configured to ignore .env to prevent accidental commits of the credentials.

3. Run the app:

   npm start
   

Sample Code for SSE Authentiocation

A simple example of showing how to implement an OAuth client provider and use it with the langchain-mcp-tools library can be found in src/sse-auth-test-client.ts.

For testing purposes, a sample MCP server with OAuth authentication support that works with the above client is provided in src/sse-auth-test-server.ts.

You can run the server with npm run sse-auth-test-server and the client with npm run sse-auth-test-client.