🧠 AI SQL ASSISTANT

Natural Language → SQL → Instant Results on Your Database

An AI-powered web application that lets you query your own databases in plain English.

Upload a .sql dump, and the app will:

  -> Create a live MySQL database
  
  -> Generate accurate SQL queries using an LLM (Hugging Face)
  
  -> Execute them instantly
  
  -> Fix errors if needed
  
  -> Display results and insights — all from a clean web UI

⚙️ Tech Stack

      Python (Flask) – Web Framework
      
      LangGraph / LangChain – Agentic Reasoning
      
      Hugging Face LLM – Query Understanding & Generation
      
      MySQL – Dynamic Database Execution

🚀 Setup & Installation

1. Clone the repo

git clone https://github.com/ashwinibhardwaj/AI-SQL-ASSISTANT.git
cd AI-SQL-ASSISTANT

2. Create virtual environment

python -m venv .venv
.venv\Scripts\activate      # Windows

or

source .venv/bin/activate   # macOS/Linux

3. Install dependencies

pip install -r requirements.txt

4 Create .env File

FLASK_APP=app.py
HUGGINGFACEHUB_ACCESS_TOCKEN=your_huggingface_api_key
MYSQL_HOST=localhost
MYSQL_USER=root
MYSQL_PASSWORD=yourpassword
MYSQL_PORT = 3306

5 Run the Application

python app.py

6 Open your browser and visit:

👉 http://127.0.0.1:5000

💻 How It Works

### Upload a .sql dump file

### Ask questions in plain English

### The AI creates a temp database, writes SQL, and executes it

### View results + natural language explanations in the UI

🌟 Key Features

🧠 AI-generated SQL Queries

⚙️ Auto Database Creation from SQL Dumps

💬 Conversational Query Interface

🔁 Error Handling & SQL Fixing

📊 Smart Result Visualization

🧩 Built on LangGraph for Agentic Workflow

👩‍💻 Author

Rudra Pratap Singh