MCP Project

This is a Model Context Protocol (MCP) project split into two main components:

Structure

• server/: Contains server-side implementations
• client_agent/: Contains client-side implementations

Prerequisites

Installing uv

This project uses uv for package management (not using anaconda). Install uv using the following commands:

# On Unix-like systems (Linux/macOS)
curl -LsSf https://astral.sh/uv/install.sh | less

# On Windows (PowerShell)
powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"
# Or using pip (cross-platform)
pip install uv

For more information about uv, visit https://github.com/astral-sh/uv

Setup

All components in this repository use uv for dependency management:

# Install dependencies for server
cd server
uv sync

# Install dependencies for client
cd ../client_agent
uv sync

Model Configuration

This repository is configured to use Ollama as the language model service running locally. To use this project, you need to:

1. Install Ollama locally on your machine from https://ollama.com
2. Pull the required model(s) using ollama pull <model-name>
3. Ensure the Ollama service is running on http://localhost:11434

The model configuration is located in client_agent/config.py where you can adjust the model settings as needed. By default, it uses gemini-3-flash-preview:cloud but you can uncomment other models as needed.

Site note

Model use in this repo is called cloud model, which is a special type of model that ollama will host on their cloud server somewhere and allow your cli to connect and get result from it

To use it

1. First you have to signup an accound on ollama.com, use this link ollama signup
2. Then process to signin on your machine

ollama signin

1. Pull your model

ollama pull gemini-3-flash-preview:cloud

Running Examples

Example 1: Math Server with Single Client

1. Start the math server:

cd server
uv run math_server.py

1. In a new terminal, run the client to test:

cd client_agent
uv run client.py

This example demonstrates communication between the math server and a single client, calculating (11 + 17) x 19.

Example 2: Multiple Servers with Multi-Server Client

1. Start both servers in separate terminals:

# Terminal 1 - Start math server
cd server
uv run math_server.py

# Terminal 2 - Start weather server
cd server
uv run weather_server.py

1. In a third terminal, run the multi-server client:

cd client_agent
uv run multi_server_mcp.py

This example demonstrates how a single client can connect to multiple MCP servers simultaneously, performing both mathematical calculations and weather queries.