Device
Browser-only transcription

Whisper Transcription

Audio is not uploaded to a server; transcription runs only in your browser.

1) Model selection
Fast loading/transcription
2) Language
3) Model setup
The first load may take time to download the model.
4) Input method
Please select an audio file like mp3/wav/m4a.
5) Start transcription
6) Results
Transcription will appear here.
Built with Next.js + Transformers.js v3 + WebGPU/CPU fallback + Web Worker · Models are downloaded from the Hugging Face Hub (CDN).