The speech is transcribed on this device.
Transcribe lectures, meetings, interviews, voice notes and videos. A speech recognition model runs on this device, so your recordings are never uploaded, and you get the text and ready-made subtitles. It is free.
How to transcribe a recording, step by step
- Drop an audio or video file on the page or press Choose a file.
- Choose the Spoken language. Choosing it is faster and more accurate than detecting it.
- Turn on Translate into English if the speech is in another language and you want English text.
- Press Transcribe. Text appears as it is written.
- Edit the transcript if you like, then press Copy text or download Text, Subtitles .srt or Subtitles .vtt.
What this tool does
This tool listens to a recording and writes down what is said. Add an audio or video file, choose the language, and press Transcribe. The text appears as it is written, so you can start reading before the whole recording is done.
Lectures, meetings and interviews
Long recordings are handled piece by piece, so an hour-long lecture works as well as a short voice note. Recordings with one clear speaker and little background noise give the best results. Put your phone close to the speaker when you record, and choose the spoken language rather than detecting it.
Subtitles for videos
Every transcript also comes as SRT and WebVTT subtitle files with timings, ready for YouTube, video editors and media players. Add your video, transcribe it, and download the subtitles.
Languages
You can choose English, Spanish, French, German, Portuguese, Italian, Dutch, Polish, Turkish, Russian, Ukrainian, Arabic, Hindi, Indonesian, Vietnamese, Tagalog, Japanese, Korean or Chinese, or let the tool detect the language. Turn on Translate into English to turn speech in another language straight into English text and subtitles.
What to do with the transcript
- Turn a lecture into study notes with notes from text.
- Make minutes and action items from a meeting with meeting minutes.
- Remove names and numbers with redact text before you share it.
- Burn the subtitles into a video with burn-in captions.
Privacy
Your recordings never leave this device. The speech recognition model is downloaded to your browser once, and all of the listening and writing happens right here. There is no upload and no server that hears your audio. The first transcription downloads the model, about 75 MB, and your browser keeps it for later.
Questions
Which files work?
MP3, M4A, WAV, OGG, Opus, FLAC and AAC audio, and MP4, MOV, WebM and MKV video. Voice memos from iPhone and Android work too.
How long does it take?
On a recent computer, a recording is usually transcribed several times faster than it plays. Phones and older computers take longer. Keep this tab open until it finishes.
Is the transcript perfect?
Speech recognition is very good on clear speech, but names, technical terms and noisy recordings can trip it up. The transcript box is editable, so you can fix anything before copying. Subtitles use the original timings.
Does it tell speakers apart?
No. It writes the words without labelling who said them. Add names by hand if you need them.