Langchain Integration

🚀 TypeScript Project: Langchain Integration with MCP Tools

This is a TypeScript project that integrates Langchain into the Model Context Protocol (MCP) tools. It enables interaction with language models and execution of various tools through a chat interface.

🚀 Quick Start

Prerequisites

• Node.js (v14 or higher)
• npm
• A locally running Ollama server
• A locally running MCP server

📦 Installation

1. Clone the repository:

git clone https://github.com/shashwat001/mcptools-langchain-integration.git
cd mcptools-langchain-integration

2. Install dependencies:

npm install

Configuration

Ollama Configuration

This project uses Ollama for LLM integration. Configure Ollama settings in src/llm.js:

export const ollamaConfig = {
    baseUrl: "http://localhost:11434",
    model: "llama3.1:8b-instruct-q6_K",
    temperature: 0.1,
    maxRetries: 2
};

MCP Server Configuration

MCP server settings can be configured in src/llm.js:

export const mcpConfig = {
    serverUrl: 'http://localhost:7000/sse',
    clientInfo: {
        name: 'ollama-client',
        version: '1.0.0'
    }
};

System Prompt

The system prompt for tool interaction can be modified in src/llm.js:

export const systemPromptForTools = "在这个环境中，你可以使用一组工具来回答用户的问题。\n 不要让用户执行函数，并自行决定是否调用工具。\n 永远不要一次调用多个工具。";

Running the Application

1. Start the Ollama server (ensure it is running on http://localhost:11434).
2. Start the MCP server (ensure it is running on http://localhost:7000).
3. Run the application:

node src/index.js

✨ Features

• LLM with an interactive chat interface
• Integration with MCP tools
• Tool execution via chat
• Support for SSE-based MCP servers

⚠️ Important Note

MCP Server Warning

This project currently uses an SSE-based MCP server. Exercise caution as the MCP server has write permissions, which may lead to unexpected system changes. Always review tool permissions before execution.