Privacy-focusedYour files are never uploaded
AI Free · Runs on your device
Turn audio into text
Transcribe interviews, lectures, voice memos and videos with Whisper, an AI speech model that runs right in your browser. Your recording never leaves your device.
or choose one from your device
Files never leave your device · Free, no sign-up · No file size limit
How to transcribe audio
- Open an audio or video file, or drop it on the page.
- Download the speech model once, choose the spoken language or leave it on automatic.
- Select Transcribe, then copy the text or download it as TXT, SRT or VTT subtitles.
Speech recognition on your device
Whisper is an open speech recognition model by OpenAI, trained on 680,000 hours of audio in about a hundred languages. Scissor Audio runs Whisper base, about 80 MB, in your browser with WebAssembly, so the recording is never sent to a transcription service.
Subtitles with timestamps
Every sentence comes with its start and end time. Download the transcript as SRT or WebVTT subtitles for YouTube, video editors and media players, or as plain text for notes and documents.
Translate speech to English
Choose Translate to English and Whisper writes an English transcript of speech in any of its languages, in one step.
Questions about Audio to text
Is my audio uploaded?
No. The audio is decoded by your browser and transcribed by a model running in your tab. Nothing is sent to a server; you can turn off your connection once the model has loaded.
Is it really free, with no limits?
Yes. The model runs on your own device, so it costs us nothing per minute: no sign-up, no minutes cap and no file size limit.
Which languages does it understand?
Whisper recognises about a hundred languages, including English, Spanish, French, German, Portuguese, Italian, Dutch, Russian, Arabic, Persian, Turkish, Hindi, Chinese and Japanese. It is most accurate in English and other widely spoken languages.
How accurate is it?
For clear speech it is close to a careful listener, though names and specialist terms may need a correction. Background music, crosstalk and distant microphones lower the accuracy.
How long does it take?
On a recent laptop a minute of audio takes roughly ten to twenty seconds. Phones and older computers are slower. The text appears as it is recognised, so you can start reading straight away.
Can I transcribe a video?
Yes. Open an MP4, MOV or WebM file and its sound is transcribed. Download the result as SRT or VTT to use it as subtitles.