xenova/whisper-web
Browser-based speech recognition powered by the Whisper transformer model running via WebGPU/Web Workers.
Not currently ranked — collecting fresh signals.
star history
This project deploys the Whisper speech recognition model directly in the browser using Hugging Face Transformers.js, enabling client-side transcription without server calls. It leverages WebGPU for GPU acceleration and Web Workers for non-blocking processing. Users can run the demo locally or via the hosted Hugging Face Space.
Frequently asked
- What is xenova/whisper-web?
- Browser-based speech recognition powered by the Whisper transformer model running via WebGPU/Web Workers.
- Is whisper-web open source?
- Yes — xenova/whisper-web is open source, released under the MIT license.
- What language is whisper-web written in?
- xenova/whisper-web is primarily written in TypeScript.
- How popular is whisper-web?
- xenova/whisper-web has 3.3k stars on GitHub.
- Where can I find whisper-web?
- xenova/whisper-web is on GitHub at https://github.com/xenova/whisper-web.