How Handy works
mainHandy provides offline speech transcription through a four-step local process:
- Trigger: Press a configurable keyboard shortcut (or use push-to-talk mode).
- Record: Speak while the shortcut is active.
- Process: Upon release, Handy processes the audio locally using either:
- Whisper models (Small, Medium, Turbo, or Large) with GPU acceleration support.
- Parakeet V3 (CPU-optimized with automatic language detection).
- Silence is filtered using Silero VAD (Voice Activity Detection).
- Output: The transcribed text is automatically pasted into your active text field.
All processing happens entirely on your local machine for privacy.