🤖 NEXA_AI — Voice-Activated AI Assistant

A voice-activated AI assistant — say the wake word and let NEXA do the rest.

🔗 Live Demo  •  GitHub

---

🎯 What is Nexa AI?

Nexa AI is a smart, voice-activated virtual assistant built with the MERN stack and powered by Google's Gemini API. Just speak your command — Nexa listens via the browser's built-in Speech-to-Text API, sends it to Gemini for an intelligent response, and replies back with a natural voice. You can also instruct Nexa to open websites, play YouTube videos, search the web, and much more — all hands-free.

✨ Features

• 🎤 Voice Activation — Say your assistant's wake word to trigger commands
• 🤖 AI Responses — Powered by Google Gemini AI for smart, natural replies
• 🔊 Text-to-Speech — Assistant speaks back in your chosen language & voice
• 🌐 Web Actions — Opens Google, YouTube, LinkedIn, Instagram, Maps & more
• 🎵 YouTube Playback — Play any song or video via voice command
• 🗓️ Real-time Info — Get current date, time, day, weather instantly
• 💬 Chat Mode — Type queries when you don't want to use voice
• 📊 Dashboard — Visual stats: total queries, voice vs chat breakdown, weekly activity
• 🕓 History Log — Browse all past interactions with search & filter
• 🧠 Custom Persona — Choose your assistant's name, avatar & personality
• 🌍 Multi-language — English, Hindi, Marathi voice support
• 🔐 Auth System — Secure register/login with JWT cookies

---

🛠️ Tech Stack

Layer	Technology
Frontend	React.js (Vite), React Router
Backend	Node.js, Express.js
Database	MongoDB (Mongoose)
AI Engine	Google Gemini API
Auth	JWT (HTTP-only cookies)
Voice	Web Speech API (SpeechRecognition + SpeechSynthesis)
Image Upload	Cloudinary
Styling	Inline styles (custom dark cyber theme)

---

📁 Project Structure

NEXA_AI/
├── Frontend/
│   ├── src/
│   │   ├── pages/
│   │   │   ├── Home.jsx          # Voice assistant main page
│   │   │   ├── Chat.jsx          # Text chat interface
│   │   │   ├── Dashboard.jsx     # Stats & analytics
│   │   │   ├── History.jsx       # Interaction log
│   │   │   ├── customize.jsx     # Assistant persona setup
│   │   │   └── Auth.jsx          # Login / Register
│   │   ├── components/
│   │   │   ├── Navbar.jsx        # Sidebar navigation
│   │   │   └── Animation.jsx     # Neural background canvas
│   │   └── contextAPI/
│   │       └── Usercontext.jsx   # Global state management
│   ├── vercel.json               # Vercel SPA + API proxy config
│   └── vite.config.js
│
└── Backend/
    ├── controller/
    │   ├── usercontroler.js      # User, history, dashboard logic
    │   └── authcontroler.js      # Register, login, logout
    ├── models/
    │   └── user.js               # MongoDB user schema
    ├── routes/
    │   ├── userRoutes.js
    │   └── authRoutes.js
    ├── middleware/
    │   ├── isAuth.js             # JWT verification
    │   └── multer.js             # File upload
    ├── config/
    │   └── cloudinary.js
    ├── gemini_api.js             # Gemini AI integration
    └── index.js

---

🚀 Getting Started

Prerequisites

• Node.js v18+
• MongoDB Atlas account
• Google Gemini API key
• Cloudinary account

1. Clone the repository

git clone https://github.com/Sakshi123509/Nexa_AI.git
cd Nexa_AI

2. Backend Setup

cd Backend
npm install

Create .env file in /Backend:

PORT=5000
MONGODB_URI=your_mongodb_connection_string
JWT_SECRET=your_jwt_secret_key
GEMINI_API_KEY=your_gemini_api_key
CLOUDINARY_CLOUD_NAME=your_cloud_name
CLOUDINARY_API_KEY=your_cloudinary_api_key
CLOUDINARY_API_SECRET=your_cloudinary_api_secret

npm run dev

3. Frontend Setup

cd Frontend
npm install
npm run dev

Create .env file in /Frontend:

VITE_API_URL=http://localhost:5000

Frontend runs on http://localhost:5173
Backend runs on http://localhost:5000

---

🎤 How to Use

1. Register → Create your account
2. Setup Persona → Go to _PERSONA and choose assistant name & avatar
3. Go to Assistant page → Click Start Listening
4. Say the wake word → e.g., "Hey Nexa, what is the weather today?"
5. NEXA responds — speaks back + opens relevant apps/links

Example Voice Commands

Command	Action
[wake word] search React tutorials	Opens Google Search
[wake word] play Kesariya on YouTube	Opens YouTube
[wake word] what time is it	Speaks current time
[wake word] open Instagram	Opens Instagram
[wake word] open maps Mumbai	Opens Google Maps
[wake word] what is machine learning	AI explanation via Gemini

---

🔑 Environment Variables

Backend

Variable	Description
MONGODB_URI	MongoDB Atlas connection string
JWT_SECRET	Secret key for JWT tokens
GEMINI_API_KEY	Google Gemini AI API key
CLOUDINARY_CLOUD_NAME	Cloudinary cloud name
CLOUDINARY_API_KEY	Cloudinary API key
CLOUDINARY_API_SECRET	Cloudinary API secret

Frontend

Variable	Description
VITE_API_URL	Backend server URL

---## 📸 Screenshots

🎙️ Assistant	⚙️ Customize
💬 Chat	🕓 History
📊 Dashboard

## 🌐 Deployment

• Frontend → Vercel
• Backend → Render
• Database → MongoDB Atlas

Vercel Config (vercel.json)

{
  "rewrites": [
    {
      "source": "/api/:path*",
      "destination": "https://nexa-ai-wy87.onrender.com/api/:path*"
    },
    {
      "source": "/(.*)",
      "destination": "/index.html"
    }
  ]
}

🤝 Contributing

Pull requests are welcome! For major changes, please open an issue first to discuss what you'd like to change.

---

Made with ❤️ by Sakshi

⭐ Star this repo if you found it helpful!