AI-powered speech-to-text application with intelligent assistant capabilities.

Currently supports ARM-based macOS only (Apple Silicon M1/M2/M3 processors)

The application is optimized for Apple Silicon Macs using Homebrew paths specific to ARM architecture. Support for Intel Macs, Windows, and Linux will be added in future releases.

• 🎤 Real-time Speech-to-Text: Convert speech to text using OpenAI Whisper with high accuracy
• 🤖 AI Assistant Integration: Activate with "Hey Sun" or "Hey Son" to get AI-generated responses using GPT-4o-mini
• 🖥️ Dynamic Overlay Interface: Seamless floating overlay that integrates with any application
• ⚙️ Persistent Settings: Encrypted configuration storage with user-friendly settings interface
• 🔄 Optimized Audio Recording: Native audio recording for ARM-based macOS systems
• 🎯 System Tray Integration: Convenient system tray access and controls
• ⌨️ Global Hotkey: Quick activation with customizable keyboard shortcuts (Cmd+H)
• 🎨 Modern UI: Clean, minimal interface with smooth animations and custom styling

Simply press the global hotkey (Cmd+H) and speak normally. Your words will be transcribed and typed where your cursor is located.

Press the hotkey and say "Hey Sun" or "Hey Son" followed by your request:

• "Hey Sun, write an email about the meeting tomorrow"
• "Hey Son, explain quantum computing in simple terms"
• "Hey Sun, create a shopping list for a dinner party"

The AI assistant will generate appropriate content and type it for you.

• Floating Overlay: Compact circular button that expands during use with visual feedback
• System Tray: Right-click for settings and controls
• Settings Panel: Configure API keys, models, and preferences with modern UI

• ARM-based macOS (Apple Silicon M1/M2/M3 processors)
• Node.js (v16 or higher)
• npm (comes with Node.js)
• Audio Recording Dependencies:
  • sox: Install via Homebrew: brew install sox
    • The app specifically looks for sox at /opt/homebrew/bin/sox (ARM Homebrew path)

1. Clone the repository:

   git clone <repository-url>
   cd SunnAI

2. Install dependencies:

   npm install
   npx electron-rebuild

3. Install sox for audio recording:

   brew install sox

4. Configure your OpenAI API key:

   • Start the application (it will prompt for API key on first run)
   • Or manually create a .env file in the root directory:

   OPENAI_API_KEY=your_openai_api_key_here
   WHISPER_MODEL=whisper-1
   ASSISTANT_MODEL=gpt-4o-mini

Development mode:

npm run dev

Production mode:

npm start

1. First Time Setup: The app will guide you through API key configuration
2. Access Settings: Right-click the system tray icon or double-click it
3. Start Dictating: Use the global hotkey (Cmd+H) and speak
4. AI Assistant: Use hotkey + "Hey Sun/Son + your command"

• macOS: Cmd+H - Start/stop recording

Build the application for distribution:

# Build for current platform (ARM macOS)
npm run build

# ARM macOS specific build
npm run build:mac

Built applications will be available in the dist/ directory.

SunnAI/
├── src/
│   ├── main.js                    # Electron main process
│   ├── services/
│   │   ├── audioService.js        # Speech recording & OpenAI integration
│   │   ├── settingsService.js     # Configuration management
│   │   └── textInjectionService.js # Text input automation
│   └── renderer/
│       ├── overlay.html           # Floating overlay interface
│       ├── overlay.js             # Overlay functionality
│       ├── overlay.css            # Overlay styling
│       ├── settings.html          # Settings panel
│       ├── settings.js            # Settings management
│       └── settings.css           # Settings panel styling
├── assets/                        # Icons and resources
├── temp/                          # Temporary audio files (auto-created)
└── dist/                          # Built applications

• Electron - Cross-platform desktop framework
• OpenAI API - Whisper (speech-to-text) & GPT-4o-mini (AI assistant)
• @hurdlegroup/robotjs - System-level text injection for macOS
• Electron Store - Encrypted settings persistence
• sox - High-quality audio recording for ARM macOS

Access these through the Settings panel:

• OpenAI API Key: Your personal API key for OpenAI services
• Whisper Model: Speech recognition model (default: whisper-1)
• Assistant Model: AI assistant model (default: gpt-4o-mini)

The application uses sox for high-quality audio recording on ARM macOS:

• Sample Rate: 16kHz
• Channels: Mono (1 channel)
• Bit Depth: 16-bit
• Format: WAV
• Path: Uses system temp directory for audio files

• Install sox: Make sure sox is installed via Homebrew: brew install sox
• ARM Homebrew Path: The app looks for sox at /opt/homebrew/bin/sox
• Microphone Permissions: Grant microphone permissions when prompted by macOS

• Ensure your OpenAI API key starts with sk-
• Check your OpenAI account has sufficient credits
• Verify internet connectivity

• macOS: Grant microphone and accessibility permissions when prompted
• Accessibility: Required for global hotkeys and text injection

• Current Support: ARM-based macOS only (Apple Silicon M1/M2/M3)
• Future Support: Intel macOS, Windows, and Linux support planned

• Intel macOS support
• Windows support
• Linux support
• Additional audio input formats
• Customizable hotkeys
• Multiple language support

MIT License - see LICENSE file for details