⚡ Velox

A blazing-fast AI chat assistant powered by Groq + Llama 3.3 70B. Featuring real-time web search capabilities, multi-turn conversation memory, and a sleek, modern dark-themed React frontend.

---

🚀 Live Demo

• Frontend: velox-henna.vercel.app
• Backend: velox-backend-6sqt.onrender.com

---

✨ Features

• ⚡ Blazing Fast: Powered by the Groq inference engine for the fastest LLM API response times.
• 🧠 Smart Context: Multi-turn conversation memory sustained per session.
• 🌐 Real-time Web Search: Native Tavily API integration for fetching up-to-date answers from the web.
• 💬 Chat History: Easily switch between multiple distinct conversations.
• 🎨 Clean UI: Beautiful, modern, dark-themed responsive frontend built with React.
• 📱 Responsive Design: Optimally designed for seamless use across mobile and desktop screens.

---

🛠️ Tech Stack

Layer	Technology
Frontend	React.js, Vite
Backend	Node.js, Express.js
LLM	Llama 3.3 70B (via Groq)
Web Search	Tavily API
Memory	Node-Cache (In-memory storage)
Frontend Deploy	Vercel
Backend Deploy	Render

---

📁 Project Structure

Velox/
├── Frontend/          # React frontend (Vite)
│   ├── src/
│   │   └── App.jsx     # Main chat UI
│   └── package.json
├── chatbot.js          # LLM agent loop with tool calling
├── server.js           # Express REST API
├── .gitignore
└── package.json

⚙️ How It Works

User Message
            │
            ▼
 Express Server (server.js)
            │
            ▼
 AI Agent Loop (chatbot.js)
            │
            ▼
   Groq API (Llama 3.3 70B)
            │
     [Needs web search?] ──► Yes ──► Tavily API ──► Results back to LLM
            │
            ▼
     Final Response ──► User

🔧 Run Locally

1. Clone the repository

git clone [https://github.com/Sakshi123509/Velox.git](https://github.com/Sakshi123509/Velox.git)
cd Velox

2. Setup the Backend

# Install root/backend dependencies
npm install

# Create a .env file in the root directory and add your keys:
# GROQ_API_KEY=your_groq_api_key
# TAVILY_API_KEY=your_tavily_api_key

# Start the Express server
node server.js

Note: The backend server will run locally on http://localhost:3001

3. Setup the Frontend

# Navigate to the frontend folder
cd Frontend

# Install frontend dependencies
npm install

# Start the development server
npm run dev

Note: The frontend application will run locally on http://localhost:5173

🌐 API Reference

Send a Message

• Endpoint: POST /chat
• Content-Type: application/json

Request Body:

{
  "message": "What is the latest news in AI?",
  "threadId": "unique-session-id"
}

Response Body:

{
  "message": "Here's the latest in AI..."
}

🔑 Get API Keys

Groq API Key: Obtain from console.groq.com Tavily API Key: Obtain from tavily.com