Audio to Text — Transcribe on Your Device
Turn a voice memo, meeting, lecture or interview into text with timestamps, right in your browser. Save it as TXT, Word or SRT/VTT subtitles. The recording is never uploaded.
Whisper on your device · timestamps · TXT DOCX SRT VTT · never uploaded
Audio to Text — Transcribe on Your Device
Every popular transcription service works the same way: you upload the recording, their servers turn it into text, and a copy of your meeting, lecture or interview now sits on someone else’s computer. This page does the transcription on your own device instead. It runs Whisper — the open speech-recognition model released by OpenAI — inside your browser tab: the recording is decoded and transcribed locally, and the text appears piece by piece with timestamps as it goes. You can copy it, or save it as a plain text file or as SRT and VTT subtitles for a video. The cost of doing it this way is a one-off download: the speech model itself (about 80 to 200 MB for Standard, 250 to 590 MB for Accurate, depending on your browser) comes from Hugging Face the first time and is then kept by your browser. Those downloads carry the model to you; nothing carries your recording away. Speed depends on your device: with WebGPU, which Chrome and Edge offer on most desktops, a minute of speech takes well under a minute; without it, expect roughly two minutes per minute of audio with the Standard model.
Never Uploaded
The recording is decoded and transcribed inside this browser tab. Only the speech model is downloaded — once — and nothing is sent back.
Timestamps and Subtitles
Every line carries its time. Save a clean .txt or Word .docx, or .srt and .vtt subtitle files that video players and editors read.
Many Languages
English, Korean, Japanese, Chinese, Spanish, German, French and more. Choose the language the recording is in.
Free, No Minutes Limit
No account, no monthly allowance of minutes. The work happens on your device, so there is no server bill to ration.
!How It Works, Step by Step
Drop a voice memo, recording or video. It is read in your browser — nothing is uploaded.
Choose the spoken language, and Standard or Accurate.
Click Transcribe. The first time, the speech model downloads once; after that it starts straight away.
Read the text as it appears, then copy it or save it as TXT, Word (.docx), SRT or VTT.
In Your Browser vs. on a Server
| Feature | Local (Us) | Cloud Tool |
|---|---|---|
| Privacy | Transcribed on your device, never uploaded | Recording uploaded and stored by the service |
| Cost | Free, no minutes limit | Free tiers capped by minutes per month |
| Speed | Depends on your device; fastest with WebGPU | Fast on their servers |
| Speaker labels | No — one continuous transcript | Often included |
Technical FAQ
How do I convert audio to text for free?
Is my recording uploaded?
How accurate is it?
How long does it take?
Why is there a download the first time?
Can I make subtitles for a video?
Does it tell who is speaking?
Which files can I transcribe?
Keep the Work on Your Own Device
Every tool here runs in your browser, with no upload and no account. Here is another one worth knowing about.
Need to cut the recording first? Use the audio editor