Skip to content
Transcribe

WAV to text, transcribed locally

Turn uncompressed WAV recordings into timed, editable text without sending them anywhere.

  • Nothing uploaded
  • No sign-up
  • Free
or drop it here · MP3 WAV M4A MP4 MKV MOV FLAC OGG
Nothing is uploaded
How it works

Three steps, all on your device

  1. Pick the WAV file

    From a field recorder, an audio interface, dictation software or an editor export. Drop it on the card or pick it.

  2. Transcribe on your device

    The browser reads the WAV directly and the Local AI engine writes down the speech, showing each line as it goes.

  3. Check and export

    Play any line against the text, fix what needs fixing, then save a document or a subtitle file.

Why it works

Why OmniaScribe

  • Raw takes stay with you

    Unedited interviews and studio sessions are transcribed without leaving your machine.

  • Every line timed

    Click a line to hear it, which makes finding the right take in a long session quick.

  • Text for your edit

    Export TXT or DOCX to plan a paper edit, or SRT to line up captions.

Why WAV files are large but simple

WAV stores sound without lossy compression, so a minute of stereo CD-quality audio takes about ten megabytes, while the same minute as MP3 needs one. Field recorders, audio interfaces and dictation tools save WAV because nothing is thrown away. For transcription the size is no problem: the file is read from your disk, not uploaded, and the speech model works on a 16 kHz mono copy that the workspace makes from it in memory.

High sample rates and multiple channels

A 48 kHz or 96 kHz WAV transcribes exactly like a 44.1 kHz one, because it is resampled first. Stereo and multichannel files are mixed down to one channel before transcription. That means a two-person interview recorded with one microphone per channel is heard as one conversation; the transcript does not say which channel a line came from.

If one channel is much louder or noisier than the other, the mix can bury the quieter voice. In that case, export each channel as its own WAV from your editor and transcribe them separately.

Using a transcript for a paper edit

Editors of documentaries, radio pieces and podcasts often cut on paper first: read the transcript, mark the lines to keep, then assemble them in the timeline. A timed transcript of the raw WAV makes this fast. Turn timestamps on, export DOCX, and highlight the lines you want. Each one carries the time it starts, so finding it again in your editor takes seconds.

Dictation recorders

Many dictation recorders and older voice recorders save WAV by default. Speak in full, natural sentences: the model adds punctuation itself and works best with ordinary speech, not spoken commands such as “comma” or “new paragraph”. Proper names, drug names and legal terms are worth checking in the transcript, and fixing them with Edit is quicker than retyping a page.

Long sessions

A recording session can run for hours. Before it starts, the workspace checks the length against what the chosen model can handle on your device and tells you if the file is too long. Split long sessions at natural breaks in your editor and transcribe each part.

Splitting has a side benefit: each part gets its own transcript in History, named after its file. Name the parts by take or by topic, and the list of transcripts becomes a rough index of the whole session. Trimming long silences and the chatter between takes before you transcribe also saves processing time, because silence still takes time to work through.

Specs

Formats and limits

Audio
MP3, WAV, M4A, AAC, FLAC, OGG, OPUS
Video
MP4, MOV, MKV, WEBM
Export
TXT, PDF, DOCX, SRT, VTT
Languages
99, or detect automatically
Longest file per model
ModelComputerPhone
Fast3 h1 h 30 min
Standard2 h1 h
High accuracy1 h30 min

Files that would take more than about 45 min on your device ask before they start.

FAQ

Questions, answered.

Something else on your mind? Email us and a person replies.

Is my WAV file uploaded?

No. The browser reads it from your device and the transcript is made there. Only the speech model is downloaded, once.

Does a 96 kHz or 24-bit WAV work?

Yes. Every WAV is converted to the format the model expects in memory, so high-resolution files work the same way.

What happens to a stereo recording?

Its channels are mixed into one before transcription. The transcript does not say which channel a line came from.

How long can the WAV be?

On a computer, Fast handles up to 3 h, Standard handles up to 2 h, High accuracy handles up to 1 h. On phones it's half of that. The length counts, not the file size.

Should I convert my WAV to MP3 first?

There is no need. WAV opens in every browser and is never uploaded, so its size does not slow anything down.

Can it say who is speaking?

Not yet. You get one stream of timed lines; add speaker names with Edit.

Guides

OmniScribe for Android

Transcribe anywhere.
No signal needed.

Install the free Android app and turn recordings into text on your phone, even with the screen off. Your audio never leaves the phone.

  • Works offline after a one-time model download
  • Translates offline, too
Get it on Google Playfree with ads · optional Pro
The OmniScribe home screen