🤖 Leo AI Assistant

Leo AI is a fully voice-controlled virtual assistant built using HTML, CSS, and JavaScript, designed for browsers and future desktop integration. It features real-time speech recognition, hotword detection, voice responses, multi-language support, and leverages Google Gemini (Gemini API) for AI-powered answers.

---

🚀 Features

• 🎙️ Voice recognition (Web Speech API)
• 🔊 Hotword detection (wake Leo by saying "Leo", "Hey Leo", etc.)
• 🗣️ Speech synthesis (AI reads answers aloud)
• 🌐 Google/Wikipedia fallback (via backend or frontend, extendable)
• 🔒 Simple login system
• 🏳️ Multi-language support (English & Hindi)
• ✨ Modern, responsive UI
• 📺 Animated avatar (video)
• 🧠 AI responses powered by Google Gemini API
• 🖥️ Ready for browser and desktop integration (via Electron, demo.exe provided)
• 👨‍💻 Easily extensible for system-level controls

---

📁 Project Structure

leo-ai/
├── logs/                   # Log files (optional)
├── node_modules/           # Installed npm packages
├── public/                 # Frontend files
│   ├── index.html          # Main UI layout
│   ├── main.js             # Frontend logic (login, voice, UI)
│   ├── style.css           # Styling
│   └── voice1.mp4          # Media asset (avatar animation)
├── .env                    # API keys and environment variables
├── demo.exe                # Compiled desktop app (optional)
├── package.json            # Project config and dependencies
├── package-lock.json       # Dependency lock file
└── server.js               # Node.js backend (AI, API, system access)

---

🛠️ Tech Stack

• Frontend: HTML, CSS, JavaScript
• Voice API: Web Speech API (speech recognition & synthesis)
• Backend: Node.js (Express)
• AI: Google Gemini API (@google/generative-ai)
• External APIs: Google Search, Wikipedia (can be extended)
• Desktop: Electron (optional, see demo.exe)

---

📸 Screenshots

---

⚡ Getting Started

✅ Without Backend (Frontend-only)

1. Clone the repo:

   git clone https://github.com/ankitbarua45/leo-ai.git
   cd leo-ai
   

2. Open public/index.html in your browser to start the assistant.

Note: Without the backend, AI answers and some features may be disabled.

---

✅ With Node.js Backend (Recommended)

To enable AI answers, system-level control, and API integration, follow these steps:

1. Install Dependencies

npm install

2. Create a .env File

In the root directory, add your Gemini API key (and others if needed):

GEMINI_API_KEY=your-gemini-api-key-here

3. Start the Server

node server.js

4. Visit http://localhost:3000 in your browser.

---

🗝️ Default Login

• Username: admin
• Password: @123

---

🧩 Extending & Customizing

• Add more wake words or languages in public/main.js.
• Integrate more APIs in server.js.
• Customize the avatar by replacing voice1.mp4.
• For desktop integration, wrap the app using Electron.

---

🤝 Credits

Developed by Ankit Barua.
Powered by Google Gemini, Web Speech API, and open web technologies.

---

📄 License

MIT License.
Copyright © 2025 Ankit Barua

---