Transcribe audio to text

FreeAudio tools

Drop an interview, lecture, meeting or podcast and get the words back with timestamps, in 20 languages. Copy the text or download TXT, SRT or VTT. Free runs each day, no sign-up needed.

  • Runs in your browser
  • Nothing uploaded
  • No watermark
Drop an audio file hereMP3, M4A, WAV, OGG, FLAC or a voice memo · up to 10 minutes per run

How to use the Transcribe Audio to Text

  1. 01

    Drop an audio file (MP3, M4A, WAV, OGG or FLAC).

  2. 02

    Pick the spoken language, or leave it on Detect automatically. For a long file, pick the 10 minutes to transcribe.

  3. 03

    Select Transcribe, then copy the text or download it as TXT, SRT or VTT.

FAQ

How long can the recording be?

Each run transcribes up to 10 minutes. For a longer file, pick the part with the slider, then run again for the next part; each part uses one free run.

How accurate is it?

Clear speech in a quiet room comes out nearly word for word. Music, crosstalk, strong accents and phone recordings lower the accuracy, and unclear words are marked [inaudible]. Read it through before you publish it.

Does it tell speakers apart?

Not yet. The transcript is split into sentences with times, without speaker names.

Is my recording stored?

Your browser cuts the part to transcribe into a small MP3 and sends only that to the speech model. SandBase does not save it. The same part within a day is answered from a copy, without using a run.