Transcribe audio to text on a Mac for free: four ways

Published 18 October 2026

You don't have to pay for transcription on a Mac. Two Apple apps already write transcripts of what they record, and two Whisper-based tools handle audio files you already have. Here's what each free transcription option on a Mac does well and where it stops.

1. Apple Notes: record and transcribe in a note

Open a note, click the Attachment button in the toolbar, choose Record Audio and click Record. Click the Transcription button to watch the text appear as you talk, and Done when you've finished. Apple says spoken words in supported languages are transcribed automatically.

Afterwards, double-click the recording and click Transcription. Clicking any line plays the audio from that point. More → Find in Transcript searches it, and More → Copy Transcript or Add Transcript to Note gets the text out. With Apple Intelligence you can also generate a summary.

The limits are on Apple's page. Transcription needs a Mac with M1 or later, and the language list is short: several variants of English, plus Spanish, Portuguese, Italian, French, German, Japanese, Korean and Chinese. Apple only describes recording inside Notes, not importing files.

2. Voice Memos: transcripts for every recording

Voice Memos on a Mac shows a transcript too, with macOS Sequoia on a Mac with Apple silicon, and Apple notes it isn't available in all countries or regions. Click the Transcription button while recording to see it live, or select a saved recording and click the same button.

Two details make it useful. The search field finds recordings by words in their transcript, not just their title. And recordings made on macOS 14 or earlier are transcribed when you open them, as long as they contain speech. So old memos become searchable without doing anything. To copy text, select it, Control-click and choose Copy.

3. whisper.cpp: any file, from the terminal

whisper.cpp is an MIT-licensed C/C++ port of OpenAI's Whisper, and on Apple silicon it runs on the GPU through Metal. It's the most flexible free option and the least friendly one. The project's quick start is:

git clone https://github.com/ggml-org/whisper.cpp.git
cd whisper.cpp
sh ./models/download-ggml-model.sh base.en
cmake -B build
cmake --build build -j --config Release
./build/bin/whisper-cli -f samples/jfk.wav

The command-line tool lists flac, mp3, ogg and wav as input formats. A Voice Memos file is m4a, so convert it first. The README's ffmpeg command produces the 16 kHz mono WAV it expects:

ffmpeg -i memo.m4a -ar 16000 -ac 1 -c:a pcm_s16le memo.wav
./build/bin/whisper-cli -m models/ggml-base.en.bin -l en -otxt -osrt -f memo.wav

-otxt saves plain text and -osrt or -ovtt saves subtitles. For non-English audio, download a multilingual model (one without .en in the name) and pass -l auto. Bigger models are more accurate and slower.

4. MacWhisper's free tier: files with a window

If you'd rather drag and drop, MacWhisper's free version covers the basics. According to macwhisper.com, it records from your Mac's microphone, opens mp3, wav, m4a, ogg, opus, mov and mp4 files, supports 100 languages and exports .txt, .srt and .vtt. It uses local models, so files stay on your Mac. Pro is a one-time €64 and adds speaker recognition, YouTube and URL transcription, batch processing, translation and more export formats.

Which one to use

• A meeting or lecture you're about to record: Notes or Voice Memos, whichever you already use.
• A file you already have, once in a while: MacWhisper's free tier.
• Many files, scripts or full control over the model: whisper.cpp.

Before trusting a transcript, play back any part where a name or number matters. Every tool on this list can mishear those.

If you want to speak new text instead

Transcription turns existing audio into text. Dictation is the other half: you talk and the words go straight into the email, note or document you're writing. Saypad does only that. Hold fn, speak, let go, and the text appears at the cursor in any app, with punctuation, recognized on your Mac. It's free with no word limit.

It won't open an audio file or record a meeting, so it doesn't replace anything above. It keeps each dictation with its audio in its history, but those are your own dictations, not imported recordings. Offline speech to text on a Mac covers both sides, and the free dictation apps compared lists the dictation options. If you also need meetings and speaker labels, how Saypad and MacWhisper differ explains who each one suits.

Questions

Can Apple Notes transcribe an audio file I already have?

Apple's guide describes transcribing audio you record inside a note. It doesn't describe importing an existing file, so for files use whisper.cpp or MacWhisper.

Do I need an Apple silicon Mac?

For Apple's transcripts, yes. Notes transcription needs an M1 or later, and Voice Memos transcripts need macOS Sequoia on Apple silicon. whisper.cpp is built from source or a package on your own Mac, and MacWhisper links its system requirements from its site.

Which free option works offline?

whisper.cpp runs entirely on your Mac once you have a model file, and MacWhisper says its local models keep files on your device. Apple's guides for Notes and Voice Memos don't say where transcription runs.

Can Saypad transcribe recordings?

No. Saypad only does live dictation: you hold a key, talk, and the text goes into the app you're using. For recordings, use one of the tools on this page.