Parrator

Simple speech-to-text tray application with global hotkey activation.

Parrator is a lightweight, always-on speech-to-text application that runs in your system tray. Press a hotkey to record speech and get instant transcription with automatic clipboard integration.

Features

• Global Hotkey Recording: Press Ctrl+Shift+; from anywhere to start/stop recording
• Automatic Transcription: Uses NVIDIA Parakeet ONNX models (AMD GPU supported) for fast, accurate speech-to-text
• Clipboard Integration: Transcriptions are automatically copied to clipboard with optional auto-paste
• System Tray: Runs silently in the background with minimal resource usage
• Cross-Platform (WIP, looking for testers):
  • Win11
  • Linux
  • Mac
• Auto-Startup: Optional system startup integration

Quick Start

Install

Simply install the .exe. Other platform executables WIP.

Prerequisites

• Python 3.11+
• Poetry for dependency management

(dev) Installation

1. Clone and install:

   git clone https://github.com/yourusername/parrator.git
   cd parrator
   poetry install

2. Run:

   poetry run python -m parrator

Usage

1. Start the app - Look for the Parrator icon in your system tray
2. Record speech - Press Ctrl+Shift+; to start recording, press again to stop
3. Get transcription - Text is automatically copied to clipboard and optionally pasted

Configuration

Right-click the tray icon and select "Settings" to edit configuration:

{
  "hotkey": "ctrl+shift+;",
  "model_name": "nemo-parakeet-tdt-0.6b-v2", 
  "auto_paste": true,
  "auto_start_with_system": false
}

Settings

• hotkey: Global keyboalllshortcut (e.g., "ctrl+shift+;", "alt+space")
• model_name: Speech recognition model to use
• auto_paste: Automatically paste transcription after copying to clipboard
• auto_start_with_system: Launch Parrator when your computer starts

Supported Models

• nemo-parakeet-tdt-0.6b-v2 (default, fast and accurate)
• nvidia/parakeet-tdt-0.6b-v2
• openai/whisper-tiny (lightweight)
• openai/whisper-base
• openai/whisper-small

System Requirements

• Windows: Windows 10/11 with DirectML support for GPU acceleration
• macOS: macOS 10.14+ (may require accessibility permissions for global hotkeys)
• Linux: Modern Linux distribution with audio support

Troubleshooting

Hotkey Not Working

• Windows: Run as administrator if needed
• macOS: Grant accessibility permissions in System Preferences
• Linux: Install python3-dev and audio libraries

No Audio Input

# Check available audio devices
python -c "import sounddevice; print(sounddevice.query_devices())"

GPU Acceleration (Windows)

• Update graphics drivers
• Install Windows Media Feature Pack
• Verify DirectML is available

Building Executable

poetry run pyinstaller Parrator.spec

Contributing

Pull requests welcome! Please ensure code follows the existing style and includes appropriate tests.