Questions

Yes. There is no account and no fee. The first transcript download is free. Another download from the same browser or network asks for your name and email so the file can be emailed to you. The transcriber runs on your own computer or phone.

No. The browser decodes the file and the Whisper model runs on your device. The only download is the model itself, the first time you use it.

The first .txt or .srt download is free and stays in your browser. A later download asks for your name and email first. After you submit the form, the file downloads in the browser and is emailed to you.

Whisper is accurate on clear speech, but no transcriber is perfect. Background noise, cross-talk, accents, and names reduce accuracy. Use the More accurate model for important recordings and proofread the result.

Up to about two hours per file. Long files take longer, roughly real time or faster on a recent laptop with the Fast model, and slower on phones.

Anything your browser can decode: MP3, WAV, M4A/AAC, OGG/Opus, FLAC, and WEBM in most browsers. Safari does not decode OGG in every version.

After the page and model have loaded once, transcription itself runs without a network connection. Reloading the page needs a connection unless your browser has cached it.

The browser downloads the Whisper model (about 40 MB for Fast, about 80 MB for More accurate) the first time. Later runs use the cached copy.

Only files on your device. Download audio only where you have the rights and the platform's terms allow it.

Yes, on recent iPhones and Android phones, but it is slower and long files can run out of memory. Use Fast and shorter files on phones.

Whisper was released as open source by OpenAI. This site is independent and not affiliated with OpenAI or Hugging Face.