Handy is a free, open-source, and extensible desktop application for offline speech-to-text transcription. It targets users seeking a privacy-focused, customizable solution for converting spoken words into text directly on their computer, without relying on cloud services. The primary benefit is local, private, and adaptable speech-to-text functionality.

How It Works

Handy utilizes a local-first architecture, processing audio entirely on the user's machine. It employs Voice Activity Detection (VAD) with Silero to filter silence, then leverages the Whisper model (specifically Whisper Small) for transcription, with support for GPU acceleration. A configurable global keyboard shortcut triggers recording, and transcribed text is automatically pasted into the active application. This approach ensures privacy and offline functionality.

Quick Start & Requirements

• Installation: Download the latest release from the project's releases page or website. Follow platform-specific installation instructions. Launch Handy and grant necessary permissions (microphone, accessibility).
• Prerequisites: While the application aims for broad compatibility, optimal performance is recommended on:
  • macOS: M-series Macs
  • Windows: Intel, AMD, or NVIDIA GPU
  • Linux: Intel, AMD, or NVIDIA GPU (tested on Ubuntu 22.04, 24.04)
• Links: Project website, GitHub Issues

Highlighted Details

• Completely offline and privacy-focused speech-to-text.
• Built with Tauri (Rust backend, React/TypeScript frontend).
• Uses Whisper Small for transcription with optional GPU acceleration.
• Includes VAD for silence filtering.

Maintenance & Community

The project is actively developed, with contributions encouraged via GitHub issues and pull requests. The project website and contact email (contact@handy.computer) are provided for community engagement. Sponsorships are acknowledged.

Licensing & Compatibility

• License: MIT License.
• Compatibility: The MIT license generally permits commercial use and linking with closed-source projects.

Limitations & Caveats

The project is actively under development, and users should consult the known issues section on GitHub for current limitations. Performance may be degraded on systems not meeting the recommended hardware specifications. Platform support is primarily focused on Apple Silicon Macs, x64 Windows, and x64 Linux.