Skip to content

ML-powered speech recognition directly in your browser

License

Notifications You must be signed in to change notification settings

Ahaxzh/whisper-web

 
 

Repository files navigation

Whisper Web

ML-powered speech recognition directly in your browser! Built with 🤗 Transformers.js.

Check out the demo site here.

Important

Experimental WebGPU support has been added to this branch (demo), if you'd like to run with GPU acceleration!

whisper-web-demo.mp4

Running locally

  1. Clone the repo and install dependencies:

    git clone https://github.com/xenova/whisper-web.git
    cd whisper-web
    npm install
  2. Run the development server:

    npm run dev

    Firefox users need to change the dom.workers.modules.enabled setting in about:config to true to enable Web Workers. Check out this issue for more details.

  3. Open the link (e.g., http://localhost:5173/) in your browser.

About

ML-powered speech recognition directly in your browser

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • TypeScript 93.0%
  • JavaScript 6.3%
  • Other 0.7%