Nova - AI Voice Assistant (JARVIS Clone)

A high-performance, hands-free, bilingual AI voice assistant powered by Google's Gemini AI and Bytez Cloud ASR. Designed to be fast, responsive, and intelligent—just like JARVIS.

Features

• Hands-Free VAD: No keys to press. Nova automatically detects when you start and stop speaking.
• 🇮🇳 Bilingual Support: Perfectly understands and speaks both Hindi and English.
• Turbo Brain: Powered by Google Gemini 2.5 Flash for near-instant thinking.
• Cloud ASR: Uses Bytez SDK with Whisper-Large-V3 for superior speech recognition accuracy.
• Premium Neural Voice: Uses edge-tts (Swara for Hindi, Ava for English) for human-like speech.
• Parallel Execution: Nova performs tasks (opening apps, etc.) while she is speaking to you.
• Direct-to-Memory: No slow temporary files (.wav or .mp3). Everything is processed in RAM for maximum speed.
• Vision: Capture photos instantly from your webcam on command.
• Turbo Typing: Near-instant text automation with support for Hindi characters.
• Smart App Control: Opens any software using Windows Start Menu search integration.

Setup

Prerequisites

• Python 3.13+
• uv (Fast Python package manager)
• Microphone & Webcam
• Google Gemini API Key
• Bytez API Key

Installation

1. Clone the repository

   git clone https://github.com/Pratham-Prog861/jarvis-clone.git
   cd jarvis-clone

2. Install dependencies

   uv sync

3. Set up Environment Variables Create a .env file in the root directory:

   GEMINI_API_KEY=your_gemini_key
   BYTEZ_API_KEY=your_bytez_key

4. Run Nova

   uv run main.py

Usage

1. Start Nova: Run the script and wait for "Nova is online".
2. Just Talk: Simply start speaking in Hindi or English.
3. Hands-Free: Nova will detect your voice, process the request, and respond automatically.

Example Commands

• Bilingual: "नमस्ते नोवा, तुम कैसी हो?" (Namaste Nova, how are you?)
• App Control: "Open Visual Studio Code" or "Chrome kholo"
• Automation: "Type 'Hello World' in Hindi" or "Likho 'Namaste Bharat'"
• Web: "Search for the latest AI news" or "Open youtube.com"
• Camera: "Capture a photo" or "Meri photo khicho"

Project Structure

jarvis-clone/
├── core/
│   ├── agent.py        # Brain (Gemini 1.5 Flash)
│   ├── emotion.py      # Sentiment Analysis
│   ├── memory.py       # Conversation History
│   └── router.py       # Action Orchestrator
├── voice/
│   ├── listener.py     # Cloud ASR (Bytez + Whisper V3)
│   └── speaker.py      # Neural TTS (Edge-TTS)
├── tools/
│   ├── browser.py      # Web & Search Tools
│   ├── camera.py       # OpenCV Photo Capture
│   ├── system.py       # Start Menu App Launcher
│   └── writer.py       # Clipboard-based Turbo Typing
├── prompts/
│   └── nova_system.txt # Personality & Logic Rules
├── main.py             # Entry Point (Parallel Loop)
└── pyproject.toml      # Modern Dependency Management

Key Dependencies

• google-genai: AI Brain
• bytez: Cloud Speech Recognition
• edge-tts: Premium Neural Voice
• opencv-python: Camera functionality
• pyautogui & pyperclip: System automation
• pygame: Audio playback (In-memory)
• sounddevice: Voice Activity Detection

License

MIT License - Feel free to use and modify!

---

Made by Pratham