⚡ IntelliSQL — Intelligent SQL Querying with LLMs

SmartBridge Generative AI with Google — Course Project

---

🎯 What Is This Project?

IntelliSQL is a web application that lets anyone query a database using plain English — no SQL knowledge required.

You type: "Show me top 5 customers by total spending" Gemini Pro writes the SQL, executes it, and explains what the results mean.

---

🔑 Why It's Unique

Feature	Description
Gemini Pro integration	Uses Google's state-of-the-art LLM to write SQL
Schema-aware prompting	Sends real table/column names to Gemini for accuracy
Result explanation	Gemini explains what the results mean in plain English
Error diagnosis	If SQL fails, Gemini tells you why and how to fix it
Editable SQL	User can edit Gemini's output and re-run it
Query history	Session-based history of all your questions
CSV export	Download results with one click
Real database	Uses SQLite with realistic e-commerce + HR sample data

---

🏗️ Architecture

User Question (English)
       │
       ▼
  Flask Backend (app.py)
       │
       ├──► GeminiHandler.generate_sql()
       │         │
       │         ▼
       │    Gemini Pro API
       │    (Natural Language → SQL)
       │         │
       │         ▼
       ├──► DatabaseManager.execute_query()
       │         │
       │         ▼
       │    SQLite Database
       │    (Runs the SQL)
       │         │
       │         ▼
       └──► GeminiHandler.explain_results()
                 │
                 ▼
            Response to Frontend
            (SQL + Results + Explanation)

---

---

🚀 How to Run

Step 1: Get Gemini API Key

Go to: https://makersuite.google.com/app/apikey Create a key and copy it.

Step 2: Set up the project

cd intellisql
pip install -r requirements.txt

Step 3: Create .env file

GEMINI_API_KEY=your_key_here

Step 4: Run the app

python app.py

Step 5: Open browser

Go to: http://localhost:5000

---

🗄️ Database Tables

customers

customer_id, name, email, country, city, age, created_at

products

product_id, name, category, price, stock, supplier

orders

order_id, customer_id, order_date, status, total_amount

order_items

item_id, order_id, product_id, quantity, unit_price

employees

employee_id, name, department, salary, hire_date, manager_id

---

💬 Sample Questions to Try

• Show all customers from India
• Top 5 products by total revenue
• Total orders per country
• Average salary by department
• Customers who spent more than 10000
• Products with stock less than 100
• Orders placed in last 90 days
• Show employees and their manager names

---

🧠 Key Concepts Used

1. Prompt Engineering — Carefully crafted prompts give Gemini the schema and rules it needs to generate accurate SQL every time.

2. Schema-Aware Context — We send the full table/column list to Gemini so it never hallucinates column names.

3. Temperature Control — Low temperature (0.1) makes Gemini deterministic, reducing random variation in SQL output.

4. Error Recovery — When SQL fails, we ask Gemini to diagnose the error and suggest a fix in plain English.

5. Safety Layer — Only SELECT queries can be executed, preventing any accidental data modification.

---

📊 Project Outcomes

After building this project you will understand:

• How to integrate Gemini Pro API into a Python app
• Prompt engineering for structured output (SQL)
• Flask REST API design
• SQLite database operations
• Session management
• Full-stack web development

---

Live Link

https://intellisql.vercel.app/