## 📍 Overview **LangGraph MCP Agents** is an experimental project that aims to transform tools defined under MCP (Modular Communication Protocol) into autonomous **agents** using the [LangGraph](https://github.com/langchain-ai/langgraph) framework. The aim of this project is to transform tools designed in accordance with the MCP protocol (e.g. tools with tasks such as planning, coding, execution, etc.) into agents running on the LangGraph infrastructure and to provide effective task sharing and collaboration between these agents. --- https://github.com/user-attachments/assets/e7c7e012-06aa-4a4e-a4cd-103141efed33 ## 📁 Project Structure ```sh └── langgraph-mcp-agents/ ├── LICENCE ├── README.md ├── app_cli.py ├── mcp_client.py ├── mcp_config.json └── requirements.txt ``` ### 📂 Project Index

LANGGRAPH-MCP-AGENTS/

__root__

mcp_client.py - MCPClient and MCPServer are the main classes in mcp_client.py
- MCPClient manages connections to multiple MCP servers, loads server configurations, starts servers, initializes tools, and handles resource cleanup
- MCPServer represents an individual server connection, manages tool execution, and handles cleanup operations
- The file is crucial for server communication and tool management in the project.

app_cli.py - App_cli.py is a command-line interface for a chatbot application that leverages the OpenAI language model
- It initializes the AI agent, accepts user input, and generates AI responses in real-time
- The script also handles server configurations, manages conversation history, and provides error handling and debugging support.

requirements.txt - Requirements.txt manages the necessary dependencies for the project
- It ensures the correct versions of libraries such as mcp, langgraph, langchain, python-dotenv, langchain-mcp-adapters, and rich are installed
- This contributes to the stability and reproducibility of the codebase across different environments.

mcp_config.json - Mcp_config.json configures the Model Context Protocol servers, specifying the commands and arguments for different server types: filesystem, SQLite, and memory
- It enables the project to interact with various data storage systems, enhancing its flexibility and adaptability to diverse environments.

LICENCE - The LICENCE file establishes the legal framework for the project, granting users the right to freely use, modify, and distribute the software under the MIT License
- It also disclaims warranties and limits liability, ensuring the software is provided "as is"
- This file is crucial for defining the terms of use and distribution of the software.

--- ## 🚀 Getting Started ### ☑️ Prerequisites Before getting started with langgraph-mcp-agents, ensure your runtime environment meets the following requirements: - **Programming Language:** Python - **Package Manager:** Pip ### ⚙️ Installation Install langgraph-mcp-agents using one of the following methods: **Build from source:** 1. Clone the langgraph-mcp-agents repository: ```sh ❯ git clone https://github.com/serkanyasr/langgraph-mcp-agents ```

LANGGRAPH-MCP-AGENTS

コメント