py-mcp-sse
by jayliangdl
This project provides a sample implementation of an MCP Client and MCP Server in Python, based on SSE. It utilizes OpenRouter.ai for LLM calls, adapting from a previous version that used ANTHROPIC LLM.
Last updated: N/A
What is py-mcp-sse?
This is an MCP (Model Context Protocol) server and client implementation using Server-Sent Events (SSE). The server provides a tool for searching books using the Gutenberg Project API, and the client can connect to and use the server's tools.
How to use py-mcp-sse?
- Clone the repository. 2. Configure the environment by installing dependencies using uv or pip and setting the OpenRouter API key in the .env file. 3. Start the server using
uv run mcp_server.py(optionally specifying host and port). 4. Start the client usinguv run client.py http://localhost:8080/sse.
Key features of py-mcp-sse
SSE-based communication
MCP client and server implementation
Gutenberg Project book search tool
OpenRouter.ai LLM integration
Decoupled server and client processes
Use cases of py-mcp-sse
Integrating LLMs with external tools
Building cloud-native applications using MCP
Creating AI assistants with access to knowledge bases
Implementing a decoupled client-server architecture for AI services
FAQ from py-mcp-sse
What Python version is required?
What Python version is required?
Python 3.13 or higher is required.
How do I configure the OpenRouter API key?
How do I configure the OpenRouter API key?
Copy the .env.sample file to .env and edit the .env file, filling in your OpenRouter API key.
How do I start the server?
How do I start the server?
Use the command uv run mcp_server.py. You can optionally specify the host and port using --host and --port.
How do I start the client?
How do I start the client?
Use the command uv run client.py http://localhost:8080/sse.
What if port 8080 is already in use?
What if port 8080 is already in use?
You can specify a different port using the --port parameter when starting the server, e.g., uv run mcp_server.py --port 8081.