Simple MCP Client Server

🚀 Simple MCP Client/Server

This project provides a simple implementation of the MCP (Model Context Protocol) client and server. It offers a quick - start guide for setting up and testing MCP - related components.

🚀 Quick Start

Prerequisites

Install uv on macOS and Linux:

$ curl -LsSf https://astral.sh/uv/install.sh | sh

Then clone the repository.

✨ Features

This project includes three main components: MCP Server, MCP Client, and MCP Observer.

📦 Installation

MCP Server (Standard Input/Output)

uv init mcp_server
cd mcp_server/
uv run main.py                        # Create a .venv environment, then delete the main.py file
uv add "mcp[cli]" httpx               # MCP server Python library
touch weather_stdio.py                # Standard input/output server implementation

MCP Client

uv init mcp_client
cd mcp_client/
uv run main.py                     # Create a virtual environment
uv add mcp
touch client_stdio.py              # MCP client implementation

MCP Observer

Use the latest versions of Node.js and npm.

curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.39.5/install.sh | bash
nvm install --lts
npm install -g npm@latest
node -v
npm -v

Install the MCP observer:

npm install -g @stellar/mcp-stardust

💻 Usage Examples

Starting the MCP Server

mcp_server$ uv run weather_stdio.py

Testing with VSCode + Cline

Add the MCP server to Cline:

Cline -> MCP Server (icon near '+') -> Installed -> Add

You'll get a JSON template "mcpServers": {}, then add a new server. Here is the cline_mcp_settings.json:

{
  "mcpServers": {
    "weather": {
            "command": "uv",
            "args": [
                "--directory",
                "/home/users/tikoehle/mcp_server",
                "run",
                "weather_stdio.py"
            ]
        }
  }
}

=> Cline displays two get_* utility functions. Click 'Done' to return to the Cline task input and try the new server.

Testing the Server

Enter tasks to call the two utilities: "Can you tell me the weather forecast for San Francisco, CA" "Can you tell me the weather alerts for San Francisco, CA" => Use my two new MCP 'weather' server utilities.

Running the MCP Client

mcp_client$ uv run client_stdio.py

Running the MCP Observer

mcp-stardust

Connecting SSE Client/Server

Connect to the SSE server using the following command:

curl --location 'http://localhost:8090/api/v1/pops?q=*<POB>' --header 'Connection: close'

Or connect via HTTP:

curl --location 'http://localhost:8090/api/v1/pops' --data-raw '{"q":"*<POB>","page":1}' 

Known Issues

Report issues in Python - SDK.

⚠️ Important Note

MCP servers developed locally should not output messages to standard output as this will interfere with protocol operations.