Qurate

⚡ AI-Powered Database Intelligence

Query your data in natural language. Get instant tables, graphs, and insights — no SQL required.

---

Qurate transforms how you interact with your databases. Ask questions in plain English, Hindi, Telugu, Punjabi, or any of 12+ languages — and get instant responses as formatted text, sortable tables, or beautiful charts.

Getting Started · Features · Architecture · API Reference · Contributing

---

🎯 Overview

Qurate is a full-stack AI platform that bridges the gap between natural language and SQL. Built for analysts, developers, and non-technical users alike, it provides:

• Conversational Database Queries — Ask questions like "Show me top 10 customers by revenue" and get structured results
• Multi-Modal Responses — Every answer is rendered as the best-fit format: text, interactive tables, or auto-generated charts
• Cross-Language Voice Input — Speak your queries in English, Hindi, Telugu, Punjabi, Tamil, Bengali, and more
• Sandboxed Execution — Each user gets an isolated PostgreSQL database, ensuring data security and privacy
• Excel-to-Database — Upload .xlsx/.csv files to instantly create queryable database tables

---

✨ Features

🧠 AI-Powered Pipeline Natural language → SQL generation via Gemini 3.1 Pro Built-in security auditor that validates every query before execution Smart response formatter that chooses text, table, or graph output Conversation memory via MongoDB for contextual follow-ups	🎨 Premium UI Awwwards-quality glassmorphic design Dark/Light mode with persistent theme toggle Framer Motion animations throughout MacBook scroll parallax, animated counters, bento grid layout
🎙️ Voice & Language Web Speech API for speech-to-text input 12+ languages: English, Hindi, Punjabi, Telugu, Tamil, Bengali, Marathi, Gujarati, Kannada, Malayalam, Urdu, French Text-to-Speech playback for assistant responses	🔐 Security & Auth JWT-based authentication with access + refresh tokens Sandboxed PostgreSQL databases per user SQL security auditing via AI before every execution Auto token refresh on 410 responses

---

🏗️ Architecture

┌─────────────────────────────────────────────────────────────────┐
│                         FRONTEND                                │
│  React 19 · Vite 8 · TailwindCSS 4 · Redux Toolkit · Recharts │
│  ┌──────────┐  ┌──────────┐  ┌──────────┐  ┌──────────────┐   │
│  │ Landing  │  │  Auth    │  │ Pricing  │  │  Dashboard   │   │
│  │  Page    │  │Login/Sign│  │  Page    │  │ Chat + Voice │   │
│  └──────────┘  └──────────┘  └──────────┘  └──────────────┘   │
│                          │ native fetch                         │
└──────────────────────────┼──────────────────────────────────────┘
                           │
                    ┌──────▼──────┐
                    │   FastAPI   │
                    │  Port 1008  │
                    └──────┬──────┘
                           │
     ┌─────────────────────┼──────────────────────┐
     │            │                │               │
┌────▼─────┐ ┌───▼────────┐ ┌─────▼──────┐ ┌─────▼──────┐
│  Redis   │ │  MongoDB   │ │ PostgreSQL │ │  Gemini AI │
│  Cache   │ │  Atlas     │ │  Sandboxes │ │  3.1 Pro   │
│          │ │            │ │            │ │            │
│ • Tokens │ │ • Users    │ │ • Per-user │ │ • SQL Gen  │
│ • User   │ │ • Chat Hist│ │   isolated │ │ • Security │
│   cache  │ │ • Accounts │ │   databases│ │ • Formatter│
└──────────┘ └────────────┘ └────────────┘ └────────────┘

Tech Stack

Layer	Technology	Purpose
Frontend	React 19, Vite 8, TypeScript 5.8	SPA with component architecture
Styling	TailwindCSS 4, Shadcn UI	Utility-first CSS with premium components
Animations	Framer Motion (motion/react)	Page transitions, micro-interactions
Charts	Recharts	Interactive data visualization
State	Redux Toolkit	Auth, chat, and UI state management
API	Native fetch wrapper	Auto token refresh, typed endpoints
Backend	Python 3.12, FastAPI	Async REST API
AI	Google Gemini 3.1 Pro	SQL generation, security audit, formatting
Auth DB	MongoDB Atlas	User accounts, chat history, sessions
Cache	Redis 7	Refresh token store, user session caching
User DB	PostgreSQL 16	Per-user sandboxed databases
Packages	pnpm (frontend), uv (backend)	Fast, disk-efficient package managers

---

🚀 Getting Started

Prerequisites

• Node.js ≥ 20 and pnpm (npm i -g pnpm)
• Python ≥ 3.12 and uv (pip install uv)
• PostgreSQL 16+ running locally
• Redis 7+ running locally (redis-server)
• MongoDB connection (Atlas or local)
• Google Gemini API Key

1. Clone the Repository

git clone https://github.com/your-username/Qurate.git
cd Qurate

2. Backend Setup

cd backend-py

# Create virtual environment and install dependencies
uv venv
uv pip install -r requirements.txt

# Configure environment
cp .env.example .env
# Edit .env with your actual credentials:
#   MONGO_URI, GEMINI_API_KEY, POSTGRES_USER, POSTGRES_PASSWORD, etc.

# Start the server
uvicorn src.main:server --port 1008 --reload

3. Frontend Setup

cd frontend

# Install dependencies
pnpm install

# Start dev server
pnpm dev

4. Open the App

Navigate to http://localhost:5173 — sign up, and you'll be redirected to the AI dashboard.

---

📁 Project Structure

Qurate/
├── frontend/                    # React SPA
│   ├── public/
│   │   └── favicon.svg          # Brand logo (purple lightning bolt)
│   ├── src/
│   │   ├── components/
│   │   │   ├── dashboard/       # Chat, Sidebar, Voice, Excel Upload
│   │   │   ├── landing/         # Hero, MacbookScroll, Features, Stats
│   │   │   ├── ui/              # GlassCard, GradientButton, ThemeToggle
│   │   │   ├── Navbar.tsx       # Responsive glassmorphic nav
│   │   │   └── ProtectedRoute.tsx
│   │   ├── lib/
│   │   │   ├── api.ts           # Native fetch wrapper with auto-refresh
│   │   │   └── utils.ts         # cn() helper
│   │   ├── pages/
│   │   │   ├── Landing.tsx      # Hero + scroll + features + stats + CTA
│   │   │   ├── Login.tsx        # Glassmorphic login form
│   │   │   ├── SignUp.tsx       # Full registration form
│   │   │   ├── Pricing.tsx      # 3-tier pricing cards
│   │   │   └── Dashboard.tsx    # AI chat interface
│   │   ├── store/
│   │   │   ├── store.ts         # Redux store config
│   │   │   ├── authSlice.ts     # Login/signup/logout
│   │   │   ├── chatSlice.ts     # Messages, prompt, history
│   │   │   └── uiSlice.ts      # Theme, sidebar, voice language
│   │   ├── types/
│   │   │   └── speech.d.ts      # Web Speech API declarations
│   │   ├── App.tsx              # Routes
│   │   ├── main.tsx             # Entry point + Provider
│   │   └── index.css            # Design system + theme variables
│   ├── package.json
│   ├── vite.config.ts
│   └── tsconfig.app.json
│
├── backend-py/                  # FastAPI server
│   ├── src/
│   │   └── main.py              # Routes: login, signup, prompt, upload-excel
│   ├── agents/
│   │   └── agent.py             # AI pipeline: SQL gen → audit → execute → format
│   ├── decorators/
│   │   └── auth.py              # JWT authentication decorator
│   ├── models/                  # Pydantic models & response helpers
│   ├── mongodb/                 # MongoDB client (users, chat history)
│   ├── redisdb/                 # Redis client (token store, session cache)
│   ├── postgresdb/              # PostgreSQL client (sandboxed DBs)
│   ├── utils/                   # JWT, password hashing
│   ├── requirements.txt
│   └── .env
│
└── README.md

---

📡 API Reference

Public Routes

Method	Endpoint	Description
POST	/sign-up	Register a new user (creates sandbox DB)
POST	/login	Authenticate and receive JWT tokens
POST	/refresh	Exchange refresh token for new access token

Authenticated Routes

All require accesstoken header

Method	Endpoint	Description
POST	/prompt2query	Send natural language → get text/table/graph
GET	/chat-history	Retrieve conversation history
DELETE	/chat-history	Clear conversation history
POST	/upload-excel	Upload .xlsx/.csv to create a DB table

Response Shapes

Every /prompt2query response includes a data field with one of three types:

// Text response
{ "type": "text", "response": "Your database has 3 tables." }

// Table response
{
  "type": "table",
  "response": {
    "columns": ["name", "age", "city"],
    "rows": [["Alice", 30, "NYC"], ["Bob", 25, "SF"]]
  }
}

// Graph response
{
  "type": "graph",
  "response": {
    "x": ["Jan", "Feb", "Mar"],
    "y": [100, 150, 200],
    "x_label": "Month",
    "y_label": "Revenue"
  }
}

---

🎨 Design System

Brand Colors

Token	Value	Usage
--q-purple	#863bff	Primary brand, CTAs, gradients
--q-purple-deep	#7e14ff	Hover states, accents
--q-cyan	#47bfff	Secondary brand, chart accents

Visual Effects

• Glassmorphism — .glass and .glass-strong utility classes
• Glow Effects — .glow-purple and .glow-cyan for hover states
• Text Gradient — .text-gradient for the brand purple→cyan gradient
• Mesh Background — .bg-gradient-mesh for ambient radial gradients
• Shimmer Loading — .shimmer for skeleton loading states

Theme Toggle

The app supports dark and light modes:

• Persisted in localStorage as qurate_theme
• Toggled via the sun/moon icon in the navbar or sidebar
• CSS variables swap automatically via the .dark class on <html>

---

🎙️ Voice Input

Qurate supports cross-language voice queries via the Web Speech API:

Language	Code	Language	Code
English	en-US	Hindi	hi-IN
Punjabi	pa-IN	Telugu	te-IN
Tamil	ta-IN	Bengali	bn-IN
Marathi	mr-IN	Gujarati	gu-IN
Kannada	kn-IN	Malayalam	ml-IN
Urdu	ur-IN	French	fr-FR

The language picker is accessible via the 🌐 button next to the microphone in the chat input.

---

🔒 Security Model

1. JWT Authentication — Short-lived access tokens (60 min) + long-lived refresh tokens (7 days)
2. Redis Token Store — Refresh tokens are stored in Redis hashes (refresh:{email}) with 7-day TTL; tokens can be individually invalidated without affecting other sessions
3. User Session Cache — Authenticated user details are cached in Redis (user:{email}) for 30 minutes, reducing MongoDB reads on every request
4. Sandboxed Databases — Each user gets an isolated PostgreSQL database; cross-DB queries are impossible
5. AI Security Audit — Every generated SQL query passes through a dedicated security agent before execution
6. No Raw SQL Exposure — Users never see or write SQL; the AI generates and the auditor validates

---

📊 Chat Response Types

The dashboard renders three distinct response types inline:

Text

Plain conversational responses, clarifications, and status messages. Includes a 🔊 speaker button for text-to-speech playback.

Table

Sortable, scrollable data tables with alternating row colors. Click column headers to sort ascending/descending.

Graph

Auto-generated bar charts using Recharts with brand gradient fills. The AI automatically selects the best chart type based on the data shape.

---

🛠️ Development

Frontend Scripts

pnpm dev          # Start Vite dev server (HMR)
pnpm build        # TypeScript check + production build
pnpm preview      # Preview production build locally

Backend Scripts

uvicorn src.main:server --port 1008 --reload    # Dev with hot reload
python start.py                                  # Production start

Environment Variables

Create a .env file in backend-py/:

MONGO_URI="mongodb+srv://..."
GEMINI_API_KEY="your-gemini-api-key"
ACCESS_TOKEN_SECRET_KEY="your-secret"
REFRESH_TOKNE_SECRET_KEY="your-refresh-secret"
ACCESS_TOKEN_DURATION="60"
REFRESH_TOKEN_DURATION="7"
POSTGRES_PASSWORD="postgres"
POSTGRES_HOST="localhost"
POSTGRES_USER="postgres"
PORT=1008

---

🤝 Contributing

Contributions are welcome! Please follow these steps:

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

---

📄 License

This project is licensed under the MIT License — see the LICENSE file for details.

---

Built with 💜 by Akshit Jain

_{Powered by Gemini AI · React · FastAPI · PostgreSQL · MongoDB}