yeabwang/demo_mcp_server

Demo MCP Server with Groq LLM Integration

This is a simple boilerplate project for running a local MCP (Model-Context Protocol) server and interacting with it using the Groq API (Llama 3, Gemma, etc.). Useful for testing tool integration, agent workflows, and LLM orchestration.

Contents

Prerequisites

• Python 3.11 or higher
• pipx and uv (for clean Python environments)
• Groq API key from https://console.groq.com
• Unix shell or Git Bash on Windows

Setup

1. Clone the repository

git clone https://github.com/yourname/demo_mcp_server.git
cd demo_mcp_server

2. Create and activate the virtual environment

uv venv
source .venv/Scripts/activate    # Windows (Git Bash)
# or
source .venv/bin/activate        # macOS/Linux

3. Install dependencies

uv pip install -r requirements.txt

Environment Variables

Create a .env file with your Groq API key and model:

GROQ_API_KEY=your-groq-api-key
GROQ_MODEL=llama3-70b-8192

Other valid model values include:

• mixtral-8x7b-32768
• gemma-7b-it

Running the MCP Server

Start the MCP server (tool defined in weather.py):

python weather.py

This exposes one tool:

@mcp.tool()
async def get_weather(location: str) -> str:
    return "Today is sunny"

It will be available at http://localhost:8000/mcp.

Running the Groq Client

This runs a test LLM prompt using your Groq credentials:

python groq_client.py

It will:

• Send a prompt to the Groq API
• Optionally connect to the MCP tool if needed

Project Structure

demo_mcp_server/
├── .env
├── weather.py
├── groq_client.py
├── main.py
├── requirements.txt
├── pyproject.toml
├── uv.lock

Notes

• You can add more tools by decorating them with @mcp.tool().