No Upload Studio
Which languages does it understand?
English, Spanish, Portuguese, French, German, Italian, Dutch, Polish, Romanian, Czech, Swedish, Russian, Ukrainian and Turkish. With Auto-detect it recognises which of them is spoken from the first words.
Transcribe Audio to Text
Turn an interview, a lecture, a voice note or a podcast into text, with the time of every line, in English and 13 other languages. The speech model runs in your browser, so the recording never leaves your device.
How to Transcribe Audio to Text
- Open your file in the audio editor and choose the Subtitles tool on the left. Transcribe is at the bottom.
- Choose the language spoken, or leave Auto-detect.
- Press Transcribe. The first time, the page says what it will download (a speech model of about 85 MB, once) and waits for Download and start.
- Wait while it listens: the progress shows how far it is, and Cancel stops it. Then click any line to play from there.
- Press Copy text, or save it: Save .txt for the text, Save .srt or Save .vtt for subtitles with the times.
Questions
Is my recording uploaded?
No. The page downloads the speech model once (your browser keeps it), and the model listens on your device. The recording and the text stay with you; nobody else hears or reads them.
How long does it take?
It depends on your device: on a recent computer, often less than the recording lasts. Long files are cut into pieces of up to half a minute at the pauses, and it works through them one by one.
How accurate is it?
Clear speech close to the microphone comes out well; music, overlapping voices and strong noise make mistakes more likely. Removing the background noise first helps. Read it through before you use it, as with any automatic transcript.
Can I turn it into subtitles for a video?
Yes: Save .srt or Save .vtt gives subtitles with these times. The video editor can also write and time subtitles for a video directly.