audio to text

Drop a recording. Get the words.

A memo from your iPhone, a WhatsApp voice message, an interview, a lecture, the sound of a video: drop it on nchova and it is transcribed on your Mac, then summed up in notes. Nothing is uploaded, no minutes are counted, and it is free with Apple’s speech model.

Try nchova free for 30 days no card, no account

macOS 26 · Apple silicon Mac

In short

Up to four hours
audio or video, one file at a time: drop several and they wait their turn.
Free, on the Mac
with Apple’s speech model, in 10 languages: no account, no upload, no limit on how many.
Then notes
the topics and what was decided, written when the file ends, and a chat to ask it questions.

Three ways in

Whichever you use, the file stays where it is: nchova reads it, and copies nothing.

  1. Choose Transcribe a File… in nchova’s menu, in the menu bar, and pick one or more files.
  2. Drop files on the Meetings window (Fn+M). A memo can come straight out of Voice Memos on the Mac.
  3. In the Finder, right-click a file and choose Open With › nchova.

What it reads

Anything the Mac itself can play, audio or video, up to four hours a file.

Audio

m4a, mp3, wav, aiff, flac, caf, aac: voice memos, recorders, podcasts, the mp3 exported from another app.

Voice messages

WhatsApp’s, saved on the Mac or sent over with AirDrop: their opus files open as they are.

Video

mp4 and mov, a recorded lecture, a call recorded by Zoom or QuickTime: nchova transcribes the sound.

Not

WebM files, films protected by DRM, and links: a YouTube video has to be a file on the Mac first.

Each file becomes a meeting

In the Meetings window, like your calls: under the day it was recorded, with its transcript, its notes and its chat.

Its title

the file’s name. When the name says nothing, like “New Recording 12” or WhatsApp’s, the Mac’s own model gives it one from what was said, with Apple Intelligence on.

Its date

when it was recorded, taken from the file itself: a memo from Tuesday goes under Tuesday.

The voices pro

told apart, Voice 1, Voice 2…, and named with a click.

The notes

the topics and what was decided, by the model you choose, as after a call. Then ask: “What did she say about the prices?”

Where it goes next

Word, PDF or Markdown, or the transcript as subtitles. Or ask Claude and ChatGPT about it, through nchova’s MCP server.

Lectures and interviews, written out

A two-hour lecture or a long interview is just a longer file: nchova reads it on the Mac, at its own pace, while you do something else. If you record a meeting or start a long dictation in the meantime, the file waits and picks up again when you are done.

With Pro, the voices are told apart, so the lecturer and the questions from the room, or you and the person you interviewed, are not one block of text. With 25 European languages from Parakeet, a lecture in Polish or Dutch is transcribed too.

Next time, you can skip the recording: with the Mac on the desk, nchova transcribes a lecture or a meeting in the room as it happens, offline.

When something else is enough

Voice Memos on the Mac transcribes its own recordings, with macOS 15 or later and Apple silicon, where Apple offers transcription. For a memo you will read once, that is enough; nchova adds the voices told apart, the notes, the chat, the export, and files that did not come from Voice Memos.

Notes records audio with a live transcript on Macs with M1 or later, in ten languages, and sums it up with Apple Intelligence. It is good for what you record in Notes; it does not take a file you already have.

WhatsApp transcribes voice messages itself, on the device: on the iPhone in twenty languages, on Android in four. On the Mac, its help page says it may not be available yet.

For a batch of files, a YouTube link, or a transcript you correct while you listen, MacWhisper is made for that.

Questions about files

00:00:04 You Is it really free?
00:00:09 nchova Yes. Transcribing files is in the free version, with Apple’s speech model, in 10 languages, with no limit on how many. Pro adds the voices told apart, Parakeet’s 25 European languages and better notes.
00:00:27 You Is my recording uploaded anywhere?
00:00:32 nchova No. The file is read and transcribed on your Mac, and stays where it was: nchova copies nothing. Only if you choose your own Claude Code or Codex to write the notes does the transcript, not the audio, go to Anthropic or OpenAI.
00:00:50 You How do I get a WhatsApp voice message onto the Mac?
00:00:55 nchova On the iPhone, share it from WhatsApp with AirDrop: it reaches the Mac as a file. Then drop it on the Meetings window.
00:01:13 You How long can a file be?
00:01:18 nchova Up to four hours. Drop several and they are transcribed one after another.
00:01:36 You Which languages?
00:01:41 nchova 29: 10 free with Apple’s model, 25 European ones with Parakeet (Pro). A recording that switches between two of your languages is transcribed in both.
00:01:59 You Can I play the audio back next to the transcript?
00:02:04 nchova Not in nchova: it keeps the words, and your file stays yours, where it was. For a transcript you correct while listening, MacWhisper is made for that.

Hold Fn.

nchova writes the rest. Thirty days of everything, free.

Download nchova