How do you transcribe an audio file on a Mac?

Start with Voice Memos if the recording is in it. It can recognize and transcribe speech in audio recordings on a supported Mac, and it comes with macOS. If it does not work for your file or your Mac, use an app that transcribes the audio on your own computer. Either way, check the transcript against the recording before you use it.

Transcribing a file is a different job from dictating as you speak. Dictation types words into a document while you talk. Transcription turns a recording you already have into text. Our question on the difference between dictation and transcription explains the split in more detail.

Voice Memos transcribes recordings, but only on supported Macs

The Voice Memos app on a Mac can recognize and transcribe the speech in audio recordings ¹. The condition that matters most is the hardware. Apple's page for macOS Sequoia says that speech in recordings can be transcribed with macOS Sequoia on a Mac with Apple silicon ². Apple also says that transcription is not available in every country or region ³.

Apple's support pages are split by macOS version, and the Apple silicon wording appears on the Sequoia page. Open the page for your own version before you assume the feature is there. If a recording shows no transcript, check three things in order: your chip, your macOS version and your region.

When Voice Memos does transcribe your file, the steps are short:

1. Open the recording in Voice Memos and view its transcription.
2. Select the text you want, Control-click it, and choose Copy ⁴.
3. Paste it into another document, such as a Mail message or a text file.

Copying one passage at a time works for short recordings. For a long interview, read the whole transcript against the audio before you use any of it.

Use a Whisper app for files that Voice Memos cannot handle

Voice Memos is built for recordings. For other files, such as an interview you downloaded or a lecture video, use an app that runs a Whisper speech model on your Mac. Whisper is a speech recognition model from OpenAI, and the apps below run speech models on the computer itself.

Buzz describes itself as transcribing and translating audio offline on a personal computer ⁵. Its documentation says it can import audio and video files and export transcripts as TXT, SRT or VTT files ⁶. The export matters if you need timestamps for subtitles. An SRT file is a plain-text subtitle format.

Buzz has a Mac requirement to check before you download it. Its README says that the Mac version now requires Apple silicon, and that 1.4.5 is the last version to support Intel Macs ⁷. If your Mac has an Intel chip, use that older release or pick a different app.

MacWhisper takes a different approach. Its website says you can transcribe any file by dragging it into the app ⁸. That is the vendor's own description, so treat it as a reason to test the app, not as proof that it handles your file. Run the test on one of your own recordings.

Whether a file stays on your computer depends on each app's settings and privacy page, not only on the fact that it runs a model locally. Read those pages before you transcribe a confidential recording.

Test one recording before you choose

Transcription accuracy depends on the audio. A quiet interview with one speaker is far easier than a noisy meeting with several people talking at once. Choose one recording that matters to you, and run the same file through each option you are considering.

Compare the results on three things that are easy to get wrong: names, numbers and technical terms. A transcript can read smoothly and still change a name or a figure, so check those words first. Count the fixes you would need to make, and compare the apps by that count rather than by how polished the text looks.

Results also vary by engine and by recording, so one test is not a verdict. Use the test to decide between the options you have, and keep a note of the recording conditions, such as a laptop microphone in a quiet room, so you can repeat the test later.

What to do

1. Open the recording in Voice Memos first, and check whether a transcript appears.
2. If it does not, pick one local app, such as Buzz or MacWhisper, and transcribe a five-minute part of the file.
3. If the audio is private, turn off Wi-Fi for that test and confirm the app still transcribes.
4. Compare the errors in names, numbers and terms, then choose the app that needs the fewest fixes.
5. Keep the original recording until you have checked the final text.

If you want an open-source option, the next question covers open-source dictation apps for Mac. For the wider choice of Mac tools, see our guide to the best dictation apps for Mac, or our overview of dictation on Mac. If your recordings are long, read how long it takes to transcribe an hour of audio before you plan your time. For the file types each tool accepts, see which audio formats can be transcribed.