QuietSteno

Home / Tools / Audio to text

Audio to text converter that keeps the file on your device

Turn any audio recording into editable text. The conversion runs in your browser, so nothing is uploaded, and there is no cap on how many files you convert.

Drop an audio or video file here

MP3, WAV, M4A, MP4, MOV, WebM, OGG or FLAC. The file stays on this device.

Runs on your device with the open-source Whisper model. Nothing is uploaded. How privacy works

To convert audio to text, open QuietSteno, choose your audio file, select the language spoken in it and wait while the speech model runs on your device. The text appears line by line with timestamps. Edit it on the page, then copy it or download a TXT, SRT or VTT file.

How to convert audio to text

  1. Add the audio fileDrag the file onto the pad or click “Choose a file”. MP3, WAV, M4A, AAC, OGG, Opus and FLAC are supported.
  2. Pick language and accuracySet the spoken language. Choose Balanced for most audio, or Most accurate when the speech is hard to follow.
  3. Read the text as it arrivesLines appear in order with a timestamp each. A progress line shows how much of the recording is done.
  4. Correct and saveFix any misheard words in place, then use Copy text or download the format you need.

Audio formats that work

QuietSteno uses your browser’s own audio decoder, so it opens the formats your browser can play:

  • MP3: the usual format for podcasts and voice recorders.
  • WAV: uncompressed audio from studio software and field recorders.
  • M4A and AAC: iPhone Voice Memos and audio-only meeting recordings.
  • OGG and Opus: common for voice messages and web recordings.
  • FLAC: lossless archives.

Every file is converted to 16 kHz mono before transcription, because that is what the speech model listens to. A high sample rate or stereo track does not improve the result; a clear voice does.

Getting cleaner text from audio

The model hears what a careful listener would hear. Three things help most:

  • Set the right language. A recording in Spanish transcribed as English produces nonsense.
  • Prefer the original file. Audio that has been re-recorded through speakers, or compressed several times, loses the consonants that distinguish similar words.
  • Step up a setting for hard audio. Most accurate copes better with accents, fast speech and background noise, at the cost of a larger download and more time.

Questions and answers

Is my audio uploaded to a server?

No. The speech model is downloaded to your browser and runs on your own processor or graphics chip. The recording is read from your disk by the page and never sent anywhere.

Is there a limit on audio length?

QuietSteno sets no limit on file count, length or minutes, and there is no account. The practical limit is your device: the browser has to hold the decoded audio in memory, so a recording of several hours can fail on a phone or an older laptop. If that happens, split the file into parts.

Can I convert a voice message or call recording?

Yes, if you can save it as a file. Voice messages are usually OGG, Opus or M4A, all of which open here. Whether you may record or transcribe a call depends on the consent rules where you live.

How accurate is the transcript?

It depends on the recording and the accuracy setting. Clear speech close to the microphone transcribes well; crosstalk, background music and strong echo cause mistakes. Names, numbers and technical terms are the most common errors, so check those against the audio by clicking the timestamps.

Does it work on a phone?

Yes, in a current mobile browser, but phones have less memory and slower chips. Use the Fast setting and shorter recordings on a phone, and switch to a laptop or desktop for long files.

Related

Last updated 13 November 2026.