aouladbouchta/sqlite-ai-agent

SQLite AI Agent 🤖🗄️

An AI-powered database assistant that lets you interact with a local SQLite database using natural language.

This project combines FastAPI, LangChain, and OpenAI models to create a smart interface where users can:

• —Browse database tables
• —View schemas
• —Preview rows
• —Query data
• —Insert records
• —Update records
• —Delete records

All directly from a web interface or chat commands.

✨ Features

AI Database Assistant

• —Natural language database queries
• —Supports SELECT / INSERT / UPDATE / DELETE

Smart Web Interface

• —AI chat interface
• —HTML table rendering
• —Chat history stored locally
• —Clear chat button

Database Explorer

• —Sidebar with all tables
• —Search tables instantly
• —Row count badges
• —View table schema
• —Preview first 20 rows
• —Auto-select table after AI operations

🧠 Example Commands

Users can type instructions like:

Show me the first 10 users
Show the schema of the orders table
Insert a new user with name John Doe
Update the status to shipped where id is 10
Delete the user where id is 7

The AI agent automatically:

1. 1.Inspects the schema
2. 2.Chooses the correct tool
3. 3.Executes the database operation
4. 4.Returns formatted results

🏗️ Architecture

User Interface
      │
      ▼
Frontend (HTML + JS)
      │
      ▼
FastAPI Backend
      │
      ▼
LangChain AI Agent
      │
      ▼
Database Tools Layer
      │
      ▼
SQLite Database

🗂️ Project Structure

sqlite-ai-agent/
│
├── app/
│   ├── main.py
│   ├── db.py
│   ├── tools.py
│   ├── agent.py
│   ├── schemas.py
│
│   ├── static/
│   │   ├── style.css
│   │   └── app.js
│
│   └── templates/
│       └── index.html
│
├── .env
├── requirements.txt
└── shop_database.db

⚙️ Installation

1️⃣ Clone the repository

git clone https://github.com/javamohammed/sqlite-ai-agent.git
cd sqlite-ai-agent

2️⃣ Create a virtual environment

python -m venv venv

Activate it:

Windows

venv\Scripts\activate

Mac / Linux

source venv/bin/activate

3️⃣ Install dependencies

pip install -r requirements.txt

4️⃣ Create .env

Create a .env file:

OPENAI_API_KEY=your_openai_api_key
DATABASE_URL=sqlite:///./shop_database.db
OPENAI_MODEL=gpt-4.1-mini

5️⃣ Add your SQLite database

Place your database file in the project root.

Example:

shop_database.db

🚀 Run the Application

Start the server:

uvicorn app.main:app --reload

Open the browser:

http://127.0.0.1:8000

🌐 API Endpoints

Chat with AI

POST /api/chat

Example request:

{
  "message": "Show me the first 10 users"
}

List Tables

GET /api/tables

Tables With Row Counts

GET /api/tables-with-counts

Table Schema

GET /api/schema/{table_name}

Table Preview

GET /api/table-preview/{table_name}

Returns the first 20 rows.

🔐 Safety Notes

This application allows direct database modification.

Operations supported:

• —INSERT
• —UPDATE
• —DELETE

To reduce risk:

• —UPDATE requires WHERE
• —DELETE requires WHERE

⚠️ Do not use this on sensitive production databases without additional protections.

Recommended improvements:

• —Authentication
• —Authorization
• —Audit logs
• —Query preview
• —Table permissions
• —Backups

⚠️ Limitations

• —Designed for SQLite
• —No built-in authentication
• —Large tables may slow down row counts
• —Natural language ambiguity may occur

🧪 Example Prompts

Show me all tables
Show the schema of users
Show me the first 20 rows from orders
Insert a new product with name Laptop and price 900
Update the status to delivered where id is 5
Delete the order where id is 2

🔧 Environment Variables

🛠️ Future Improvements

Possible next features:

• —Pagination
• —Column sorting
• —Row filtering
• —SQL preview
• —CSV export
• —Authentication
• —Multi-database support
• —Dark mode
• —Query history

🤝 Contributing

Contributions are welcome.

Steps:

1. 1.Fork the repository
2. 2.Create a branch
3. 3.Commit your changes
4. 4.Open a pull request

📄 License

You can use the MIT License for this project.

Example:

MIT License
Copyright (c) 2026

🙏 Acknowledgements

FastAPI https://fastapi.tiangolo.com/

LangChain https://www.langchain.com/

OpenAI https://platform.openai.com/

SQLAlchemy https://www.sqlalchemy.org/