Transcribe a conversation as it happens.

Is it happening now, or already recorded?

These are two different jobs. Live, the words appear while people talk, and the transcript is ready the moment the conversation ends. With a recording, a tool works through an audio file afterwards.

For a file, Transcribe in Word takes .wav, .mp4, .m4a and .mp3 uploads. Microsoft lists 300 minutes of uploaded audio a month with a Microsoft 365 subscription, and Word saves the audio to OneDrive. Free open-source tools can also transcribe a file on your own computer with nothing uploaded, and the guide linked at the foot of this page walks through the setup. The rest of this page is about the live case.

What your phone or computer can already do

On an iPhone, open a note, tap the attachment button, then Record Audio, then the record button. Tap the transcription button to watch the words appear. Apple lists this for iPhone 12 and later when the phone's language is English, Spanish, Portuguese, Italian, French, German, Japanese, Korean, or Simplified or Traditional Chinese. For a phone call, tap the More button during the call, then Call Recording. Apple asks you to make sure the other person is willing to be recorded, and both of you hear a notice.

On Android, Live Transcribe shows the words as people talk, and Google says it works without internet in languages you have downloaded. On a Pixel, the Recorder app keeps the audio with its transcript, and Google says the transcription happens on the phone. Pixel 6 and later can label speakers.

A phone is the easy choice away from your desk. A computer is the better fit when the call is already on it, in any app. You also get a big screen to read along on, and the text is already on the computer you write on.

How to transcribe a conversation on a Mac or PC with Fraze

1. Install Fraze

   Download it for a Mac with Apple silicon on macOS 14 or later, or for Windows 11. No account is needed.

2. Choose what it listens to

   Click the dark audio pill. Check Microphone for people in the room with you, and pick the microphone closest to them from the list under it. Check System audio for the other side of a call, or both for both sides, which the pill shows as Mix.

3. Pick the spoken language

   Click the language pill, set Spoken language, and set Translate to No translation. The first time, Fraze downloads the language files. After that it needs no internet.

4. Ask, then start

   Ask the people you are talking with before you begin. Then press Start listening. Words appear while people talk, and each line settles a moment after the speaker pauses.

5. Keep the text

   When you stop listening, the session is saved in History on your computer, where you can copy it or save it as a plain text file.

Tips for a cleaner transcript

• Put the microphone between the people talking. For a group around a table, an external microphone in the middle hears everyone more evenly than a laptop at one end.
• On a call with both sources on, wear headphones, so the microphone picks up only your voice.
• Take turns where you can. Overlapping voices are hard to write down, by hand or by app.
• Add names and unusual words to Custom vocabulary in Settings, one per line. It applies when the next session starts.
• Lines break where someone pauses, but nothing marks who spoke. Add names as you read it back, while the conversation is fresh.
• Fraze keeps no audio. If you will need to check an exact quote later, record on your phone as well, with everyone's agreement.

Why it runs on your computer

A talk with a doctor, a teacher or a source is not something everyone wants sent to a server. Fraze turns speech into text on your computer and saves the transcript there. Nothing is uploaded, no bot joins a call, and Fraze never saves the audio. If you want no copy at all, turn off Keep transcripts on this computer in Settings.

That also means it works on a plane, in a basement meeting room or with Wi-Fi off, once the language files are downloaded.

What to expect, honestly

• It is live only. It cannot open an audio or video file. You can play a recording on your computer and let it listen, but that takes as long as the recording.
• There are no speaker labels, and the saved text has no timestamps. History keeps plain lines.
• There is a short delay. Lines settle about half a second to a second after the speaker pauses.
• It hears 25 European languages on Mac and Windows, among them English, Spanish, French, German, Italian, Portuguese, Dutch, Polish and Ukrainian. It also hears Chinese, Japanese, Korean, Cantonese, Vietnamese and Thai on Mac and Windows.
• It needs macOS 14 or later on Apple silicon, or Windows 11 on x64 with a 4-core AVX2 processor and 8 GB of memory. There is no phone app.

What it costs

The first 30 minutes are free, with no account and no card. After that it is $29.99 once or $2.99 a month, both unlimited. Because the work happens on your computer, there is no per-minute charge, so a two-hour interview costs the same as a short chat.

Questions