Transcription — on your own machine

Your recordings are never uploaded anywhere.

Your machine does the transcription; only the text reaches the server. Then you search for a word and get the minute it was said.

Take a look →

No per-minute fee. No upload. No queue.

After an hour of recording, you will never listen back.

You record the meeting because "it might come in handy". Then forty recordings sit there, and when you need that one sentence — what was promised, and by when — you do not listen through it. You ask again instead.

0:16 I can reach the internet specialist.

2:04 The internet will work there too.

Search gives you the recording first, then the minute. Inflected forms are matched too.

0:03 Not to make a ~~repair vidio~~ repair video, obviously.

On request it reviews the text — and shows what it changed. The raw transcript stays.

Three steps, then forget about it

Drop the recording in a folder Voice memo, meeting, dictated note. The machine watches the folder.
It transcribes — locally A 50-second recording takes 18 seconds on a MacBook Pro. The audio never moves.
Searchable within minutes Open it in the browser, search, done.

What I commit to

—The audio file never reaches the server. Not a setting — a consequence of the architecture.
—Correction never starts on its own. You decide when, and on which recording.
—Every processing step is logged: what left, when, under which settings.
—The raw transcript stays intact. You can rewrite any line, and it shows that you did.

And what I do not

This is a tool, not a finished box. Transcription needs a machine of yours running the program, and it has to be set up. There is no sign-up and no mobile app.

In exchange there is no per-minute fee, and your recordings stay with you.

The words were always there. They just could not be found.

Take a look, and search for a word in one of your own recordings.