Transcribing used to mean paying a service and uploading your audio. On a current Mac it doesn’t have to: Apple’s speech recognition now runs on the Mac itself, it’s good at long recordings like lectures and meetings, and it’s free. This guide covers every way to turn audio into text on macOS 26 and 27, from the built-in apps to subtitle files, and how to get the most accurate result.
Which method should you use?
| You want to… | Use | Free | Offline | Existing files | Timestamps or SRT |
|---|---|---|---|---|---|
| Record a meeting or lecture and read along | Notes | Yes | Yes, after a one-time download | Partly | No |
| Transcribe a voice memo | Voice Memos | Yes | Yes, after a one-time download | M4A recordings | No |
| Type by speaking | Dictation | Yes | On Apple silicon, for supported languages | No | No |
| Transcribe any audio or video file | A transcription app | Varies | Yes, with on-device models | Yes | Yes |
| Get subtitles for a video | A transcription app, Whisper | Varies | Yes | Yes | Yes |
The built-in transcription in Notes and Voice Memos needs a Mac with Apple silicon (M1 or later). It doesn’t need Apple Intelligence; that’s only required for the summaries those apps can also write.
Method 1: Notes, for recording with a live transcript
Notes can record audio and transcribe it as you speak, which makes it great for lectures, interviews and meetings in the room.
- Open a note and click the Attachment button (the paperclip) in the toolbar, then choose Record Audio.
- Click Record. Click Transcription to watch the text appear live. You can pause and resume.
- When you’re finished, click Pause, then Done.
- To read it later, double-click the recording and choose Transcription. Click any word to play the audio from that point.
- Use More › Copy Transcript to paste the text elsewhere, or Add Transcript to Note to put it in the note itself. Find in Transcript searches it.
Supported languages include English (US, UK, Canada, Australia, New Zealand and South Africa), Spanish, French, German, Italian, Portuguese, Japanese, Korean and Chinese. The first time, Notes downloads the speech model, which needs an internet connection; after that it works offline.
You can also drag an existing audio file, such as an M4A, MP3 or WAV, into a note and open it to get a transcript, although Apple doesn’t document this, so results can vary by file.
Method 2: Voice Memos, for recordings you already have
Voice Memos on the Mac transcribes on macOS 15 and later with Apple silicon.
- Open Voice Memos and select a recording (or record a new one).
- Click the Transcription button to show the text. During a new recording, click it to watch the transcript live.
- To copy the text, select it, Control-click and choose Copy, or use Copy Transcript from the menu.
- Use the search field to find a recording by any word that was said in it.
You can bring an existing recording in by dragging it into Voice Memos. It reliably accepts M4A; if an MP3 or WAV won’t import, convert it first in QuickTime Player with File › Export As › Audio Only, or Control-click it in Finder and choose Services › Encode Selected Audio Files. For noisy recordings, try the Enhance Recording option before transcribing.
Method 3: Dictation and Live Captions (live speech only)
Dictation types what you say into any text field. Turn it on in System Settings › Keyboard › Dictation, then press the Microphone key, or the shortcut you choose (such as pressing Fn twice), and speak. On Apple silicon, supported languages are processed on the Mac.
Live Captions (System Settings › Accessibility › Live Captions) shows on-screen captions for any audio playing on your Mac, including calls and videos.
Neither one can open a file, and Live Captions doesn’t save what it shows, so they’re no help for transcribing a recording you already have. A common workaround, playing a file out loud into Dictation, is slow and inaccurate. Use a method that reads files directly instead.
Method 4: A transcription app, for any file and for subtitles
The built-in apps are fine for notes, but they don’t give you timestamps, subtitle files or easy batch work, and they’re fussy about file types. For those you need an app that transcribes files directly.
Since macOS 26, Apple gives apps the same on-device speech model that powers Notes and Voice Memos, called SpeechAnalyzer. Apple describes it as built for long-form and distant audio, such as lectures, meetings and conversations, and it produces precise timings for every word, which is what you need for subtitles.
MightyBench’s Transcribe uses that model. Drop in MP3, M4A, WAV, AIFF, MOV or MP4 files, choose the language, and save as plain Text, SRT or VTT subtitles with timings, or Timestamped text. Each language is a one-time download from Apple, after which it works offline, and nothing is uploaded. SRT files can go straight to Burn In Subtitles to put captions into the video itself.
Free and open source: Whisper
Whisper is a speech recognition model released as open source by OpenAI, and it supports around 100 languages. On a Mac, the easiest ways to use it are:
- whisper.cpp, a free command-line version tuned for Apple silicon. Install it with
brew install whisper-cpp, download a model, and run it on a 16 kHz WAV file. It can write TXT, SRT and VTT files. Larger models are more accurate but slower: the “small” model is a reasonable balance. - MacWhisper, a Mac app built on Whisper with a free tier and a paid Pro version.
If you’re comfortable in Terminal, whisper.cpp is free and very capable. The trade-off is setup: you need to convert audio to WAV first (for example with FFmpeg), choose and download models, and remember the flags.
What about Microsoft Word’s Transcribe?
Word’s Transcribe feature works in Word for the web, not in the Word app for Mac. It needs a Microsoft 365 subscription, uploads your audio to Microsoft’s servers, and has a monthly limit on uploaded minutes. It’s an option if your recordings can go to the cloud; it isn’t offline.
How to get a more accurate transcript
Speech recognition is very good now, but the recording matters more than the software.
- Get the microphone close. A phone on the table in front of the speaker beats a laptop across the room.
- Reduce background noise. Close windows, turn off fans, and avoid noisy cafés. Clean Up Audio can remove steady background noise before you transcribe.
- One speaker at a time. Overlapping voices are the hardest thing for any transcription tool. Apple’s built-in transcripts don’t label speakers.
- Choose the right language and region. Transcribing British English with a US model, or a mixed-language recording with the wrong language set, lowers accuracy noticeably.
- Proofread names and jargon. Product names, acronyms and people’s names are where every model slips. In our own testing, Apple’s model heard “macOS” as something quite different.
- For video, extract the audio first if a tool won’t open the video. Extract Audio saves the soundtrack as M4A or WAV.
Privacy: where does your audio go?
| Runs on your Mac | Sends audio to the cloud |
|---|---|
| Notes and Voice Memos transcription | Microsoft Word Transcribe |
| Dictation on Apple silicon (supported languages) | Online transcription websites |
| Live Captions | Cloud options in some transcription apps |
| Apps using Apple’s SpeechAnalyzer model | Google Docs voice typing |
| whisper.cpp and other local Whisper apps |
The on-device options need an internet connection once, to download the language model. Notes may also ask whether you want to share recordings with Apple to improve transcription; that’s optional.
For interviews, medical or legal recordings, client calls and anything confidential, an on-device method avoids sending the recording to a third party at all.
Frequently asked questions
Can a Mac transcribe audio to text for free?
Yes. On an Apple silicon Mac with macOS 15 or later, the Notes and Voice Memos apps transcribe recordings for free, on the Mac. Whisper-based tools such as whisper.cpp are free as well and can transcribe any audio file.
Does Voice Memos transcribe on Mac?
Yes, on macOS 15 or later with an Apple silicon Mac. Select a recording and click the Transcription button. You can copy the text or search your recordings by what was said.
Does Notes transcription need Apple Intelligence?
No. Transcription in Notes needs an M1 or later Mac and a supported language. Apple Intelligence is only needed for the summary feature.
How do I transcribe an existing audio file on a Mac?
Drag the file into Voice Memos (M4A works best) or into a note, then open its transcript. For MP3s, videos, batches or subtitle files, use a transcription app that opens files directly, such as one built on Apple’s SpeechAnalyzer model or on Whisper.
Can I transcribe audio to SRT subtitles on a Mac?
Not with Notes or Voice Memos, which give plain text without timings. Use a transcription app that exports SRT or VTT, such as MightyBench’s Transcribe or MacWhisper, or whisper.cpp in Terminal.
Is transcription on a Mac private?
Notes, Voice Memos and apps using Apple’s on-device speech model transcribe on your Mac after a one-time language download, so the audio isn’t uploaded. Web services and Microsoft Word’s Transcribe send the audio to their servers.
Does Microsoft Word for Mac have Transcribe?
Not in the desktop app. Microsoft offers Transcribe in Word for the web, which works on a Mac in a browser with a Microsoft 365 subscription and an internet connection.
How accurate is Mac transcription?
With a clear recording of one person speaking a supported language, accuracy is high enough to need only light proofreading. Background noise, distant microphones, overlapping speakers, accents and specialist vocabulary all lower it, so check names and technical terms.