🚀 LeepSQL-AI

Natural Language to SQL with Enterprise-Grade Security

Query your database in plain English. AI generates SQL. Multiple security layers protect your data.

---

🎯 What is LeepSQL?

LeepSQL-AI lets non-technical users query databases using natural language while maintaining enterprise security through:

• 🤖 AI-Powered SQL Generation - Type "Show me top 10 customers" → Get SQL
• 🛡️ 5-Layer Security - AI validation + Backend checks + Human approval
• 👁️ Human-in-the-Loop - Review SQL before execution
• 🔒 Read-Only by Default - No accidental data modifications

---

🏗️ Architecture

┌──────────────┐     ┌──────────────┐     ┌──────────────┐
│   Frontend   │────▶│   Backend    │────▶│  AI Agents   │
│   (React)    │     │  (FastAPI)   │     │ (LangGraph)  │
└──────────────┘     └──────┬───────┘     └──────────────┘
                            │
                            ▼
                     ┌──────────────┐
                     │   Database   │
                     │ (PostgreSQL) │
                     └──────────────┘

Key Principle: AI Agent NEVER accesses database directly. Backend is the single control plane.

---

📁 Project Structure

LeepSQL-AI/
├── frontend/          # React + TypeScript + Vite + Tailwind
├── backend/           # FastAPI + SQLAlchemy + Pydantic
├── agents/            # LangGraph AI Pipelines
│   ├── generator_pipeline/   # NL → SQL
│   ├── evaluator_pipeline/   # Security Check
│   └── executer_pipeline/    # SQL Execution
└── Docs/              # Detailed Documentation

---

🚀 Quick Start

Prerequisites

• Python 3.10+
• Node.js 18+
• PostgreSQL database
• Ollama with Qwen model

1. Start AI Agents

cd agents
pip install langchain langgraph langchain-ollama psycopg2-binary fastapi uvicorn
python server.py

2. Start Backend

cd backend
pip install -r requirements.txt
python -m uvicorn app.main:app --reload --port 8000

3. Start Frontend

cd frontend
npm install
npm run dev

4. Open Browser

http://localhost:5173

---

🔐 Security Layers

Layer	Component	Protection
1	Frontend	Input validation
2	Backend API	Pydantic schema validation
3	AI Evaluator	LLM security analysis
4	Backend SQL Check	Keyword blocking (DROP, DELETE, etc.)
5	Human Approval	Manual review before execution
6	Execution	Timeouts + Row limits

---

💡 Example Usage

User Types:

"Show me all customers from New York who spent more than $1000"

AI Generates:

SELECT * FROM customers 
WHERE city = 'New York' AND total_spent > 1000
LIMIT 1000

User Approves → Results Displayed

---

🛠️ Tech Stack

Layer	Technology
Frontend	React 18, TypeScript, Vite, Tailwind CSS
Backend	FastAPI, SQLAlchemy, Pydantic, psycopg2
AI	LangGraph, LangChain, Ollama, Qwen3-VL
Database	PostgreSQL (MySQL support planned)

---

📡 API Endpoints

Backend

Method	Endpoint	Description
POST	/db/connect	Connect to database
POST	/query	Process NL query
POST	/approve	Approve pending SQL

AI Agents

Method	Endpoint	Description
POST	/api/generate	NL → SQL
POST	/api/evaluater	Security check
POST	/api/executer	Execute SQL

---

📖 Documentation

Detailed docs available in /Docs:

• Project Overview
• Backend Architecture
• Frontend Guide
• AI Agents Deep Dive
• Security Model
• API Reference
• FAQ & Cross Questions

---

🔒 Security Features

• ✅ Credentials stored in memory only (never logged)
• ✅ Multi-layer SQL validation
• ✅ Blocked operations: DROP, TRUNCATE, ALTER, INSERT, UPDATE
• ✅ DELETE only with WHERE clause
• ✅ Query timeout (30s) and row limit (1000)
• ✅ Human approval mode

---

🤝 Contributing

1. Fork the repository
2. Create feature branch (git checkout -b feature/amazing)
3. Commit changes (git commit -m 'Add amazing feature')
4. Push to branch (git push origin feature/amazing)
5. Open Pull Request

---

📄 License

MIT License - See LICENSE file for details

---

👥 Team

Built for GDG Hackathon - Democratizing data access with AI safety.

---

LeepSQL-AI - Query smarter, not harder 🚀