MCP Client-Server Python Example

What is MCP Client-Server Python Example?

This project demonstrates a simple client-server implementation of the Model Context Protocol (MCP), an open protocol by Anthropic that enables large language models to interact with external tools and resources. It is intended for developers learning how to build MCP servers and clients using Python.

How to use MCP Client-Server Python Example?

Install dependencies with uv sync, create a .env file with OPENAI_API_KEY and MCP_SSE_URL, start the server with uv run server.py --host 0.0.0.0 --port 8080, then run the client with uv run client.py. The client will list available tools and start an interactive chat loop. Type quit to exit.

Key features of MCP Client-Server Python Example

• Exposes tools (e.g., addition) and resources (e.g., greetings) via SSE.
• Client connects to server and lists available tools.
• Uses OpenAI GPT models to process queries and call tools.
• Built with Python 3.12+ and the MCP Python SDK.
• Uses SSE (Server-Sent Events) for transport.
• Managed with uv for fast dependency resolution.

Use cases of MCP Client-Server Python Example

• Learning how to build an MCP server and client from scratch.
• Prototyping an agentic workflow where an LLM calls external tools.
• Testing SSE-based communication between MCP components.
• Demonstrating integration of MCP with OpenAI’s chat completions API.
• Educational reference for implementing the Model Context Protocol.

FAQ from MCP Client-Server Python Example

What is MCP?

MCP stands for Model Context Protocol, an open protocol introduced by Anthropic that allows language models to interact with external tools, APIs, and resources in a standardized and secure way. It supports agentic workflows and multi-channel communication.

What are the runtime requirements?

Python 3.12 or higher, along with the MCP Python SDK, OpenAI Python SDK, Uvicorn, python-dotenv, and uv as the package installer.

How do I set up the server?

Create a .env file with an OPENAI_API_KEY and the server SSE URL (MCP_SSE_URL=http://localhost:8080/sse), then run uv run server.py --host 0.0.0.0 --port 8080. The server exposes tools and resources at the /sse endpoint.

How do I run the client?

In a separate terminal, run uv run client.py. The client will connect to the server, list available tools, and start an interactive prompt. Type your query, and the client uses OpenAI to decide whether to call a server tool.

How do I stop the client?

Type quit at the prompt to exit the interactive chat loop.