🚀 MCP Server & Database Integration using Python

A Python-based Model Context Protocol (MCP) server built using FastMCP that exposes database operations as MCP tools. The project integrates a local SQLite database and demonstrates how AI clients such as Claude Desktop and a custom Python client can interact with backend services through MCP.

---

📌 Project Overview

This project was developed as part of my internship to understand and implement the Model Context Protocol (MCP).

The server exposes database operations through MCP tools, allowing AI applications and clients to interact with a structured SQLite database.

The project demonstrates:

• Building an MCP server using FastMCP
• Integrating a SQLite database with multiple tables
• Exposing database operations as MCP tools
• Connecting through a Python client
• Integrating with Claude Desktop
• Automated testing using pytest

---

✨ Features

• ✅ Python-based MCP Server using FastMCP
• ✅ SQLite Database Integration
• ✅ Users, Products and Orders Database
• ✅ CRUD Operations
• ✅ Search Functionality
• ✅ Parameterized SQL Queries
• ✅ Structured JSON Responses
• ✅ Logging
• ✅ Python MCP Client
• ✅ Claude Desktop Integration
• ✅ Automated Testing using pytest

---

🛠 Tech Stack

Technology	Purpose
Python 3.13	Programming Language
FastMCP	MCP Server Framework
SQLite	Local Database
pytest	Automated Testing
Claude Desktop	MCP Client
MCP Inspector	Tool Testing
VS Code	Development Environment

---

🏗 System Architecture

                 Claude Desktop
                       │
                       │
                (MCP Tool Calls)
                       │
                       ▼
            ┌─────────────────────┐
            │    FastMCP Server   │
            │      server.py      │
            └─────────────────────┘
                  │         │
                  ▼         ▼
             MCP Tools   Logging
                  │
                  ▼
             Database Layer
                 db.py
                  │
                  ▼
            SQLite Database
      (Users • Products • Orders)

---

📂 Project Structure

mcp-python-server/
│
├── client.py              # Python MCP client
├── server.py              # FastMCP server exposing database tools
├── db.py                  # Database access layer
├── seed.py                # Creates and seeds the SQLite database
├── requirements.txt
├── pytest.ini
├── README.md
│
├── tests/
│   ├── test_db.py         # Database tests
│   └── test_tools.py      # MCP tool tests
│
└── .gitignore

---

🗄 Database Schema

The project uses a SQLite database consisting of three tables.

Users

Column	Type
id	INTEGER
name	TEXT
email	TEXT

Products

Column	Type
id	INTEGER
name	TEXT
category	TEXT
price	REAL
stock	INTEGER

Orders

Column	Type
id	INTEGER
user_id	INTEGER
product_id	INTEGER
quantity	INTEGER
order_date	TEXT

---

🔧 MCP Tools

User Tools

Tool	Description
get_all_users	Retrieve all users
get_user	Retrieve a user by ID
search_user	Search users by name
add_user	Add a new user
remove_user	Delete a user

---

Product Tools

Tool	Description
get_all_products	Retrieve all products
get_product	Retrieve a product by ID
search_product	Search products

---

Order Tools

Tool	Description
get_all_orders	Retrieve all orders

---

🚀 Installation & Setup

Clone the repository

git clone https://github.com/<your-username>/mcp-python-server.git

cd mcp-python-server

Create a virtual environment

python -m venv venv

Activate the environment

venv\Scripts\activate

Install dependencies

pip install -r requirements.txt

Generate the SQLite database

python seed.py

---

▶ Running the Project

Start the MCP Server

mcp dev server.py

Run the Python Client

python client.py

Execute Tests

pytest

Expected Output

===========================
10 passed
===========================

---

📸 Screenshots

MCP Inspector

---

Claude Desktop Integration

---

Python Client

---

Automated Testing

---

🤖 Claude Desktop Integration

The MCP server can be integrated with Claude Desktop using the MCP configuration file.

Once configured, Claude Desktop can directly invoke the available MCP tools.

Example prompts:

• Show all users.
• List all products.
• Search for Laptop.
• Show all orders.
• Create a new user.

---

🧪 Testing

Automated testing was implemented using pytest.

The test suite validates:

• Database operations
• CRUD functionality
• Search operations
• Invalid user handling
• Invalid product handling
• MCP tool responses

Current Status

✅ 10 Tests Passed

---