🎙️ maVoice

🚀 Open-Source Voice Dictation Powered by Groq's Lightning-Fast Inference

Experience the future of voice-to-text with Groq DEV Tier - Ultra-fast transcription that leaves OpenAI's free tier in the dust!

┌─────────────────┐
│  🎤 maVoice     │  ← Tiny floating widget (100x22px)
│ ▶ ■ ▪ ▪ ▪ ▪    │    Always on top of your screen
└─────────────────┘    Double-click to start!

---

✨ Features

• ⚡ Blazing Fast: Powered by Groq's Whisper Large v3 Turbo model - the fastest inference in the game
• 🎯 Native Performance: Built with Rust and Tauri for minimal resource usage
• 🎨 Beautiful UI: Sleek, modern floating widget that stays out of your way
• 🔒 Privacy First: Your API key, your data - everything stays local
• 🌐 Cross-Platform: Works on Linux (Windows and macOS coming soon!)
• 🎤 Smart Recording: Real-time audio visualization and voice detection
• 📋 Instant Copy: Automatic clipboard integration for seamless workflow
• ⚙️ Advanced Settings: Comprehensive configuration panel with model selection
• 🎛️ Intuitive Controls: Double-click to start, single-click to stop
• 🌍 Multi-Language: Support for 100+ languages with custom prompts

🎯 What is maVoice?

maVoice is a floating voice dictation widget that lives on your desktop. Unlike traditional apps with windows and menus, maVoice is a tiny, always-accessible button that floats above your other applications.

The Floating Widget Design

Normal State           Recording            Processing           Success
┌─────────────┐       ┌─────────────┐     ┌─────────────┐     ┌─────────────┐
│ 🎤 maVoice  │  →    │ 🔴 ▶▶▶▶     │  →  │ 🟠 ◈◈◈◈◈    │  →  │ ✅ Done!    │
└─────────────┘       └─────────────┘     └─────────────┘     └─────────────┘
   (Blue)                 (Red)              (Orange)            (Green)

• Size: 100x22 pixels (compact floating button)
• Behavior: Always on top, transparent background, no window borders
• Dragging: Right-click or Ctrl+Left-click to drag to a new position

🏎️ Why Groq DEV Tier?

Feature	Groq DEV Tier	OpenAI Free
Speed	🚀 Lightning Fast	🐌 Slow
Rate Limits	💪 400 RPM	😔 Limited
Model	🧠 Whisper v3 Turbo	🤖 Basic Whisper
Quality	🎯 Premium	📉 Variable

🚀 Quick Start

🌟 ONE-COMMAND Install

# Clone and install everything automatically
git clone https://github.com/lliWcWill/maVoice-Linux.git
cd maVoice-Linux
./install.sh

# Add your Groq API key
echo "VITE_GROQ_API_KEY=your_groq_api_key_here" > src-tauri/aquavoice-frontend/.env

# Launch!
npm run dev

Prerequisites

• Node.js 18+
• Rust 1.70+
• A Groq API key (Get one here)

Platform-Specific Setup

🪟 WSL2 Setup (Windows Users)

✨ BREAKTHROUGH: WSL2 + WSLg provides PERFECT voice dictation with zero audio issues!

Prerequisites

1. Update WSL2 (from Windows PowerShell as Administrator):

   wsl --update
   wsl --version  # Ensure version 2 with WSLg

2. Install Debian/Ubuntu if you don't have it:

   wsl --install -d Debian

Installation

# Install Rust
curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh
source $HOME/.cargo/env

# Install system dependencies
sudo apt update && sudo apt install -y \
    build-essential pkg-config libgtk-3-dev libwebkit2gtk-4.1-dev \
    libsoup-3.0-dev libjavascriptcoregtk-4.1-dev libdbus-1-dev \
    libappindicator3-dev librsvg2-dev libasound2-dev \
    xdotool wl-clipboard wtype

# Clone and run
git clone https://github.com/lliWcWill/maVoice-Linux.git
cd maVoice-Linux
./install.sh

🐧 Native Linux Setup

Debian/Ubuntu:

sudo apt update
sudo apt install -y \
    build-essential pkg-config libgtk-3-dev libwebkit2gtk-4.1-dev \
    libsoup-3.0-dev libjavascriptcoregtk-4.1-dev libdbus-1-dev \
    libappindicator3-dev librsvg2-dev libasound2-dev \
    xdotool wl-clipboard wtype

Fedora/Arch - See detailed instructions

📦 Build Debian Package

# Build the .deb package
npm run build

# The .deb file will be in:
# src-tauri/target/release/bundle/deb/

🎮 Usage

Desktop App

1. Launch maVoice - The app appears as a sleek floating widget
2. Double-click to start - The microphone activates with visual feedback
3. Speak naturally - Real-time audio visualization shows your voice
4. Single-click to stop - Transcription appears instantly
5. Copy & paste - Text is automatically copied to clipboard

Web Interface (http://localhost:5173)

• Settings panel - Click the gear icon for full configuration
• API key setup - Secure local storage of your Groq key
• Model selection - Choose from Whisper variants
• Custom prompts - Add technical terms, names, or style instructions
• Temperature control - Adjust creativity vs accuracy
• Multi-language - Support for 100+ languages

Keyboard Shortcuts

• Ctrl+, - Open settings
• Alt+Space - Toggle recording
• Double Alt - Quick record
• Spacebar - Stop recording (while active)

🛠️ Tech Stack

🤝 Contributing

We love contributions! Whether it's:

• 🐛 Bug reports
• 💡 Feature requests
• 🔧 Pull requests
• 📖 Documentation improvements

Check out our Contributing Guide to get started.

📈 Performance

maVoice leverages Groq's incredible inference speed:

• Transcription Speed: < 500ms for 30-second audio
• Memory Usage: < 50MB idle, < 100MB active
• CPU Usage: < 5% during transcription
• Network: Minimal bandwidth usage with smart chunking

🔐 Privacy & Security

• Local First: All processing happens on your machine
• No Telemetry: We don't track anything
• Secure API: Your Groq API key is stored locally and never shared
• Open Source: Audit the code yourself!

📜 License

maVoice is MIT licensed. See LICENSE for details.

🙏 Acknowledgments

• Groq - For providing insanely fast inference
• Whisper - OpenAI's amazing speech recognition model
• Tauri - For making native apps actually enjoyable to build
• You - For choosing open-source!

---

Built with ❤️ by developers who were tired of slow dictation

maVoice - Where speed meets simplicity