MCP SQL Server with GPT Chat Interface

This project consists of two main components: a MCP server for handling SQLite database operations via Ollama, the easiest way to get up and running with large language models, and a web application for interacting with the database through a chat interface.

Screenshots

Project Structure

.
├── server/         # Backend SQLite server
└── gpt-db-chat/   # Frontend React application

Server

The server component is a TypeScript-based backend application that:

• Handles SQLite database operations
• Provides a REST API for database interactions
• Includes database initialization scripts

Setup

First, install Ollama and download the model qwen2.5:latest

ollama run qwen2.5:latest

Feel free to install and use other models

Setup server

cd server
npm install

npm run init:db # Initalize the database

Key files:

• server.ts - Main server implementation
• init-db.ts - Database initialization script

GPT DB Chat

The application provides a chat interface for interacting with the SQLite database.

Setup

cd gpt-db-chat
npm install

Key directories:

• app/ - Core application code
• app/chat/ - Chat interface implementation
• app/components/ - Reusable UI components
• app/infra/ - Infrastructure code including API clients

Run apps

1. Start the server:

cd server
npm run start

2. Start the web application:

cd gpt-db-chat
npm run dev

Tech Stack

• TypeScript
• Node.js
• SQLite
• React
• Vite
• React Router