lokeshtalamala1/mcp-project

MCP Project - Database Interaction

This project demonstrates a Model Context Protocol (MCP) server and client implementation for interacting with a PostgreSQL database.

Setup Instructions

Follow these steps to set up and run the project:

1. Install dependencies manager uv:

pip install uv

2. Create a virtual environment and activate it:

uv venv
.venv/Scripts/activate

3. Install project dependencies from requirements.txt:

uv add -r requirements.txt

Running the Server

To run the server with the MCP Inspector for development:

uv run mcp dev server/mcp_server.py

To run the server normally (without inspector):

uv run server/mcp_server.py

Running the Client

To interact with the MCP server:

uv run server/mcp_client.py

Example Usage

Once the client is running, you can ask natural language queries such as:

Provide me the last 6 months transactions of a customer CUST_000023

Project Structure

• .env : Environment variables (e.g., API keys, database URL).
• .gitignore : Git ignore rules.
• .python-version : Python version used.
• mcp_server.py : Main MCP server implementation.
• mcp_client.py : MCP client for interacting with the server.
• pyproject.toml : Project configuration and dependencies.
• requirements.txt : Required Python dependencies.
• server.json : MCP server configuration.
• uv.lock : Lock file generated by uv.

Workflow

1. mcp_client.py loads server.json → launches MCP server (mcp_server.py) using uv
2. MCPClient connects to MCPServer
3. User types query → MCPAgent uses OpenAI + memory + MCP tools
4. If tool get_customer_transactions is needed → Queries PostgreSQL → Formats results → Returns response
5. Chatbot prints the result