Skip to content

Speech Recognition

chrome-only alpine-badge uno-badge mit

free, accurate and unlimited speech recognition web app powered by Google's Web Speech API

🚀 launch app

Motivation

In around 2016, the built-in voice recognition support for Cantonese in both Windows and macOS were not available yet. Even for Mandarin, the quality of recognition in both OS were far below Google. So I created this little webapp, harnessing Google web speech API, to serve my daily voice recognition need.

To speed up the operation, I added auto copy to clipboard and shortcut key feature afterwards. It then became a handy always-on utility on my computer.

Now, in 2022, the built-in voice recognition for both OS are mature enough that I can rely on them over 90% of time. However, the recognition quality of Google is still much better, especially in Cantonese. So I still come back to this tool when the OS fail to recognize those less common phases or slang.

Features

demo 🎬 https://youtu.be/D2NwsPozwFw

  • auto copy recognition results to clipboard
  • multilingual
  • hotkey: space bar to toggle start / stop recognition
  • PWA: installable on desktop / mobile like native app

Usage

  • grant microphone and clipboard permission in Chrome
  • space bar: toggle start / stop recognition
  • recognition results will be displayed and copied to the clipboard

Tech Details

🔗 source code

It started with a single HTML file with internal css and javascript. As the project evolve I want to keep it simple and don't want to mess with bundler. This lead to the following tech stack selection:

Platform

It uses Google's webkitSpeechRecognition api so it only works on Chrome OS / browser.

Need Help?

Open a github issue or ping me on Twitter