Handy is a speech-to-text application whose creation was motivated by a personal experience: the developer broke a finger, resulting in his hand being out of commission, and found that existing applications were neither open source nor extensible. Handy was created to fill this gap.

Project Goals and Philosophy

Handy is not trying to be the best speech-to-text app. It is trying to be the most forkable one.

Handy is intended to be a starting point for those who need accessibility features, want to experiment with voice computing, or simply prefer tools they can actually own and modify. The project is intentionally tiny and extensible. The philosophy is to encourage the community to take it, fork it, modify it, break it, fix it. The objective is to serve as both a useful tool and a foundation for others to build upon, allowing the community to create something better than any single person could build alone. The user is invited to make Handy perfect for them.

Features and Usage

Handy is described as probably the most simple speech-to-text app. Its only function is to put whatever you say into a text box.

Usage involves a keyboard shortcut:

1. Press and hold a keyboard shortcut.

2. Speak.

3. Release.

Your words then appear wherever you were typing. Recording can be started/stopped via a shortcut or by using the "push-to-talk" mode.

Technical Details ("Under the Hood")

Handy is designed for maximum autonomy and flexibility:

• Offline Operation: It runs completely offline using the Whisper model.

• Transcriptions: It transcribes speech using Whisper Small.

• Silence Filtering: Silence is removed via a VAD filter (silero).

• Integration: The text is pasted directly into whatever app you are using.

• Resources: It runs entirely on your machine and benefits from GPU acceleration when available.

• Accessibility: It requires no subscriptions or cloud services.

• Compatibility: Handy works on Windows, macOS, and Linux.

Contributions and Community

The creator is actively looking for contributors who want to help make the codebase cleaner, more robust, and more accessible. The hope is that Handy can be the simplest starter codebase for someone wanting to play around with a desktop speech-to-text application.

Users are encouraged to implement their own features. It is suggested that this might be possible without any programming experience with the help of Claude Code and agentic programming tools. If you love Rust and see places where the code can be better, you are invited to submit PRs (Pull Requests) or reach out at contact@handy.computer, with the goal of collectively learning from a well-patterned simple codebase.