AI Free · Runs on your device
Record your voice, get the text
Free, and your audio stays on your device. Record from your microphone, listen back, save the file and turn your speech into text with on-device AI.
Text never leaves your device · Free, no sign-up · No length limit
Voice recorder
How to record your voice and transcribe it
- Select Record and allow the microphone when your browser asks.
- Pause and resume as you like, then select Stop and listen back.
- Download the recording, or select Transcribe to turn it into text you can copy, edit or polish with the AI tools.
Your recording stays with you
The browser records the microphone into memory on this device. Nothing is streamed or uploaded: the recording exists only in this tab until you save it, and it is gone when you close the page or select Delete.
Speech to text with on-device AI
Transcription uses Whisper base, an open speech recognition model that understands about 100 languages. Scissor Text downloads it once from its own server, about 80 MB, and runs it in your browser with WebAssembly.
Questions about Voice recorder
Is my recording uploaded?
No. Recording, playback, saving and transcription all happen in your browser tab. The page has no server that could receive audio, and you can check in your browser's developer tools that nothing is sent.
Is it free, with no time limit?
Yes. There is no sign-up, no subscription and no recording limit. Long recordings take longer to transcribe, since the work runs on your own device.
Which formats can I save?
The format your browser records in, which is small: WebM or Ogg in most browsers and M4A in Safari. Or WAV, which is larger and opens in any audio app.
Why does my browser ask for the microphone?
A web page can only hear the microphone after you allow it. If you blocked it by mistake, allow the microphone for text.scissor.studio in your browser's site settings and try again.
Which languages can it transcribe?
About 100, including English, Spanish, French, German, Chinese, Arabic, Hindi and Japanese. It detects the language on its own, or you can choose it for better results.