cjpais/Handy

What it solves

Handy is a privacy-focused, open-source speech-to-text application that allows users to transcribe speech into text directly into any active text field on their computer. It eliminates the need for cloud-based transcription services, ensuring that voice data remains local and private.

How it works

Users trigger recording via a configurable keyboard shortcut (hold, toggle, or tap). The application uses Silero VAD (Voice Activity Detection) to filter silence and processes the audio using local AI models. Users can choose between various Whisper models (Small, Medium, Turbo, Large) with GPU acceleration or the CPU-optimized Parakeet V3 model. Once processed, the transcribed text is pasted directly into the active application.

Who it’s for

It is designed for users seeking a free, private, and cross-platform (Windows, macOS, Linux) transcription tool for accessibility or general productivity, as well as developers who want an extensible speech-to-text framework.

Highlights

• Fully Offline: All transcription happens on the local machine without cloud connectivity.
• Extensible: Open-source and supports custom Whisper GGML models.
• Cross-Platform: Native support for Windows, macOS, and Linux.
• Flexible Control: Supports keyboard shortcuts, CLI flags for remote control, and a Raycast extension for macOS.
• Model Choice: Offers a variety of Whisper and Parakeet models to balance performance and hardware capabilities.