Handy is a free and open source speech to text application that runs on Windows, macOS, and Linux.

I built it because I broke my finger and was put in a cast, which made doing my day job really difficult trying to type with just one hand. I tried applications like MacWhisper and SuperWhisper, but I really needed something more extensible so I could play with accessibility features for my particular situation. I also love open source software and wanted to contribute something that other people can build on top of.

The app is pretty simple – you hold down a keyboard shortcut while speaking (or press once to start/stop recording), and it transcribes what you said and pastes it directly into whatever text field you're using. It's built on whisper.cpp, runs completely offline with GPU acceleration, and doesn't require any subscriptions or cloud services.

This is my first Rust project, and since I built it mostly one-handed, a lot of it was written with help from LLMs. The code definitely shows it, there are bugs and rough edges.

What I'm really hoping for from the HN community is Rust contributors who want to help fix bugs and clean up the codebase. The goal is to make this a solid foundation that people can fork and extend for their own needs, whether that's accessibility features, voice computing experiments, or just having speech-to-text software you actually own and control. I do hope that Handy can be a nice simple, clean, code base, for speech-to-text that is easily extensible for people to try what's next. And I really would love your help in making that possible.

If you've been looking for something like this, give it a try. And if you're a Rustacean who wants to help make it better, I'd love the help, or any advice.

GitHub: https://github.com/cjpais/Handy