MCP Sse Langchain Example

🚀 Multi-Server MCP Demonstration with SSE Transmission

This project demonstrates how to use the Model Context Protocol (MCP) with multiple servers, employing different transmission methods (standard input/output and Server-Sent Events).

This project is built upon the following example: https://github.com/langchain-ai/langchain-mcp-adapters

🚀 Quick Start

System Requirements

• Python 3.8+
• OpenAI API key

Installation

1. Clone the repository:

   git clone https://github.com/yourusername/mcp-sse.git
   cd mcp-sse
   

2. Install the required dependencies:

   pip install -r requirements.txt
   

3. Set the OpenAI API key in the .env file:

   OPENAI_API_KEY=your-api-key-here
   

Usage

Run the main application:

python main.py

This will:

1. Start the weather server on port 8000.
2. Connect to the math and weather servers.
3. Run the agent and demonstrate the processing of two example queries.

✨ Features

• A math server providing basic arithmetic operations (addition, multiplication).
• A weather server providing simulated weather information, supporting queries for different locations.
• A main application that connects to two servers and communicates using MultiServerMCPClient.
• Integration of LangChain and OpenAI to create an agent capable of invoking tools from two servers.

📦 Installation

1. Clone the repository:

   git clone https://github.com/yourusername/mcp-sse.git
   cd mcp-sse
   

2. Install the required dependencies:

   pip install -r requirements.txt
   

3. Set the OpenAI API key in the .env file:

   OPENAI_API_KEY=your-api-key-here
   

📚 Documentation

Project Structure

• main.py: The main application that connects to two servers and runs the agent.
• math_server.py: A simple MCP server providing arithmetic operation services using standard input/output transmission.
• weather_server.py: A simple MCP server providing weather information services using SSE transmission.
• requirements.txt: List of Python dependencies.
• .env: Environment variable file (containing the OpenAI API key).

How It Works

The application demonstrates how to use MultiServerMCPClient to connect to multiple MCP servers using different transmission protocols:

1. The math server uses standard input/output (stdio) transmission, a simple communication method based on pipes.
2. The weather server uses Server-Sent Events (SSE) transmission, a mechanism for servers to push notifications to clients in the HTTP protocol.

The main application:

1. Starts the weather server as an independent process.
2. Connects to two servers using MultiServerMCPClient.
3. Creates a LangChain agent capable of invoking tools provided by two servers.
4. Demonstrates the agent's ability to perform arithmetic calculations and obtain weather information.

Usage

Run the main application:

python main.py

This will:

1. Start the weather server on port 8000.
2. Connect to the math and weather servers.
3. Run the agent and demonstrate the processing of two example queries.

Example Queries

This demonstration includes the following two example queries:

1. Math query: "what's (3 + 5) x 12?"
2. Weather query: "what is the weather in nyc?"

Project Expansion

You can expand the project in the following ways:

1. Add more functions to the math or weather server.
2. Create other MCP servers with different functions.
3. Modify the agent to handle more complex queries.
4. Use a real weather API instead of simulated data.

📄 License

MIT

Acknowledgments

• This project uses the Model Context Protocol (MCP) developed by Anthropic.
• It integrates LangChain for agent functionality.