Guides / Recording lectures, podcasts and interviews
How to transcribe audio on a Mac for free
Tested on macOS 26.4.1 (Tahoe), Apple M3 Max, 3 Sept 2026. Updated 2026-09-03.
A Mac can transcribe audio for free in four ways. Voice Memos and Notes transcribe their own recordings on macOS 15 or later with Apple silicon. MacWhisper's free tier and the open-source Whisper model transcribe any existing WAV, MP3 or M4A file offline. macOS Dictation types live speech but stops after 30 seconds of silence and is not built for files. Record first, then transcribe.
Can a Mac transcribe an audio file for free?
Yes, and the right tool depends on where the audio comes from.
- Voice Memos: Apple states that "With macOS 15 or later and a Mac with Apple silicon, speech in your audio recordings can be recognized and transcribed to text". It transcribes recordings made in the app, live while you record or afterwards, and it adds that "Audio transcription isn't available in all countries or regions".
- Notes: on macOS Sequoia 15 or Tahoe 26 with an M1 or later Mac, Notes records audio into a note and produces a searchable transcript. Apple lists English, Spanish, Portuguese, Italian, French, German, Japanese, Korean, Simplified Chinese and Traditional Chinese. With Apple Intelligence on, it can summarize the recording.
- MacWhisper: a Mac app around OpenAI's Whisper model. The free tier transcribes MP3, WAV, M4A, OGG, OPUS, MOV and MP4 in 100 languages with speaker recognition, entirely on the Mac. Pro is 64 euros one-off.
- Whisper itself: open source under the MIT license, installed with
pip install -U openai-whisper. Models range from tiny (39 M parameters, about 10x faster than large) to large (1,550 M parameters).
The catch with the Apple tools is that they transcribe what they recorded. For a file from a lecture stream, a call or a podcast you need MacWhisper, Whisper, or an online service.
How do I transcribe a voice recording on Mac with Voice Memos?
Requirements: macOS 15 or later, Apple silicon, a supported region. Tested on macOS 26.4.1:
- Open Voice Memos and record, or select an existing recording made in the app. Apple notes that a recording made on macOS 14 or earlier is transcribed automatically when opened.
- Click the transcription button. While recording, the transcript appears live with the current word highlighted; afterwards, it shows the full text.
- Select the text, Control-click, choose Copy, and paste into Notes, Pages or an email.
- To find a passage later, type a phrase into the Voice Memos search field; results include recordings whose transcript contains it.
Two limits. Voice Memos records the microphone, so it can transcribe a lecture in the room or your own dictated notes, but it cannot record system audio, so a Zoom lecture or a podcast episode never gets in. And there is no import of arbitrary files documented for the Mac app. For those cases, go to the next section.
How do I transcribe an existing audio file on Mac?
For a WAV, MP3 or M4A already on disk, MacWhisper is the shortest route and it runs offline.
- Download MacWhisper and open it. The free tier is enough for plain transcripts.
- Drag the audio file into the window. On first use it downloads a Whisper model; larger models are more accurate and slower.
- Pick the language or leave it on automatic. Turn on speaker recognition for an interview.
- Export as TXT, SRT or VTT. Pro adds DOCX, PDF and Markdown.
From the Terminal, the reference implementation does the same: whisper recording.mp3 --model turbo after installing with pip. The GitHub page lists the turbo model at 809 M parameters and about 8x the speed of large, needing around 6 GB of memory. Online, Otter.ai imports files up to 5 GB in MP3, M4A, WAV and several video formats, but the Basic free plan transcribes only the first 30 minutes of each file, allows three imports per account and 300 minutes per month. A 90-minute lecture therefore needs a paid Otter plan or a local tool. Keep files at MP3 320 or 16-bit WAV; both transcribe equally well and MP3 is a fifth of the size.
Does macOS Dictation transcribe audio files?
No. Dictation converts live speech from the microphone into text at the cursor. Apple's guide says "You can dictate text of any length without a timeout. Dictation stops automatically when no speech is detected for 30 seconds." That makes it good for dictating notes and unsuitable for a file: any pause in the recording ends the session, and Dictation has no way to read a file at all.
The workaround that circulates online is to route the Mac's playback into Dictation through a virtual audio driver such as BlackHole set as the input. It works badly. Dictation expects one speaker, close to a mic, and it inserts punctuation as it goes; a lecture played through it comes back as a wall of text with the 30-second silence cut-off waiting at every pause. Whisper-based tools handle the same file in a fraction of the time with paragraph breaks and, in MacWhisper, speaker labels.
The workflow that holds up is record first, transcribe second. Record the source at a known format, then hand the file to the transcriber. For anything the Mac plays, a system audio recorder is the first step; AirCheck records it through ScreenCaptureKit and exports MP3 320 that MacWhisper opens directly, as in how to record a lecture on a Mac. Meeting notes have their own file-size rules, covered in recording audio for AI meeting notes.
You can dictate text of any length without a timeout. Dictation stops automatically when no speech is detected for 30 seconds.
Apple, macOS User Guide, Use Dictation on Mac
Free ways to transcribe audio on a Mac (September 2026)
| Tool | Transcribes an existing file | Runs offline | Requirement | Length limit | Price |
|---|---|---|---|---|---|
| Voice Memos | no (own recordings only) | yes | macOS 15+, Apple silicon | none documented | free |
| Notes (Record Audio) | no (own recordings only) | yes | macOS 15+ or 26, M1 or later | none documented | free |
| macOS Dictation | no (live mic only) | depends on setting | any Mac | stops after 30 s of silence | free |
| MacWhisper (free tier) | yes: MP3, WAV, M4A, OGG, OPUS, MOV, MP4 | yes | macOS app | none documented | free, Pro 64 euros |
| Whisper (open source) | yes | yes | Python, up to about 10 GB memory for large | none | free, MIT license |
| Otter.ai Basic | yes: up to 5 GB per file | no | account | 30 min per file, 3 imports, 300 min per month | free |
Requirements and limits quoted from the vendor pages listed in sources.
Where AirCheck fits
AirCheck ($39, macOS 14 or newer) is the recording half of the workflow, not a transcriber. It captures what the Mac plays through ScreenCaptureKit or any mic and interface, and exports MP3 320 or WAV that MacWhisper and Whisper accept directly. That matters when the audio is a stream, a call or a podcast that Voice Memos and Notes cannot hear, and when the file should stay on the Mac for privacy.
When you do not need it: If you are recording your own voice or a lecture in the room on a Mac with Apple silicon and macOS 15 or later, Voice Memos records and transcribes in one app for free. If the file already exists, MacWhisper alone is enough.
AirCheck records what your Mac plays, or any input, to one file. $39 once.
macOS 14 or newer. One-off payment. Fourteen-day refund, no questions asked.
Questions people ask
How to transcribe audio to text for free on Mac?
For your own recordings, Voice Memos or Notes on macOS 15 or later with Apple silicon. For any existing file, MacWhisper's free tier or the open-source Whisper model, both offline. Otter.ai's free plan works online but caps each file at 30 minutes.
How to transcribe audio from a video on Mac?
MacWhisper opens MOV and MP4 directly, so no extraction is needed. For other tools, extract the audio track first; see extracting audio from a video file on a Mac.
How to transcribe an audio file to text in Word on Mac?
If you would rather not depend on Word's own Transcribe feature, transcribe with MacWhisper, export TXT (or DOCX in Pro), and open that file in Word. The audio then never leaves the Mac.
Which model should I pick in MacWhisper or Whisper?
Start with turbo or small. Whisper's own table lists tiny at about 10x the speed of large and turbo at about 8x. Use large only for difficult audio or rare languages; on speech in a quiet room the smaller models are usually accurate enough.
Sources
- Apple: View a transcription of a recording in Voice Memos on Mac
- Apple: Record and transcribe audio in Notes on Mac
- Apple: Use Dictation on Mac
- MacWhisper
- OpenAI: Whisper on GitHub (models, license)
- Otter.ai: Import an audio or video file
- Otter.ai: Conversation, import, and app limits on the Basic free plan
Related guides
- How to record a lecture on a Mac
- How to record audio for AI meeting notes on a Mac
- How to record an interview on a Mac
- Can Voice Memos on a Mac record system audio?
- How to record a podcast on a Mac
- How to record an online course or training on a Mac
- Recording lectures, podcasts and interviews: all guides
- Every guide, by topic