File transcription versus live dictation.
These two are usually compared and rarely compete. One writes out a recording, the other types while you speak. Here is which one you actually want.
| Feature | Zumbo | MacWhisper |
|---|---|---|
| What it is for | Dictation, live, into any app | Transcribing files you already have |
| Price | $15 once, $20 for three Macs, $79 for ten | Free version, Pro €64 once |
| Runs offline | Yes | Yes |
| Types into the focused app | Yes, in about a fifth of a second | - |
| Transcribe an existing recording | No | Yes |
| Live meeting transcription | Yes, through the microphone | Recordings, with speaker recognition |
| Speaker labels | Yes | Speaker recognition on recordings |
| Custom vocabulary | 21 packs, plus your own words | - |
| Learns from corrections | Yes, Teach a word once | - |
| Notes and reminders | Yes, say a time and it becomes a reminder | No |
| Lives in the notch | Yes | No |
| Account required | No, ever | No account |
| Languages | 25 European languages, English first | - |
| macOS minimum | macOS 14 or later, Apple silicon | - |
| Download size | About 570 MB | - |
Read off macwhisper.com on 21 September 2026. A dash means the vendor does not state it on its own public pages, not that the answer is no.
What it is for
This is the whole page in one row. MacWhisper takes a recording and writes it out. Zumbo puts words where your cursor is while you talk. They are different jobs and most people who need one do not need the other.
Price
Both are bought once, which is the right model for software that runs on your machine. MacWhisper also has a free version, which is a good way to find out whether file transcription is what you actually wanted.
Runs offline
Both do the work locally. Neither uploads your audio to transcribe it, which is why either is a reasonable answer in an environment that cannot send recordings out.
Types into the focused app
Zumbo's entire design is that the text arrives where the cursor already was, fast enough that you keep talking. MacWhisper's output is a transcript you then do something with.
Transcribe an existing recording
Zumbo does not take a file. If you have an hour of recorded interview on disk, MacWhisper is the answer and this page will not pretend otherwise.
Live meeting transcription
Zumbo transcribes the room as the meeting happens, a timestamp for each paragraph. On a call, use the speakers rather than headphones and the other side is picked up too. MacWhisper's speaker recognition works on recordings.
Speaker labels
Both can tell speakers apart. The difference is when: Zumbo labels them while the meeting is running, MacWhisper after you have a file.
Custom vocabulary
MacWhisper does not state a vocabulary feature on the pages checked. Zumbo's packs cover fields like developer tools, legal, medical, design and marketing, switched on by what you say you dictate most.
Learns from corrections
Select the word that came out wrong, type what you meant, press Teach, and it is heard your way from then on. The correction stays on your Mac.
Notes and reminders
Press New note and talk without pasting anywhere. Say remind me to call Dana tomorrow at ten and there is a reminder for tomorrow at ten.
Lives in the notch
Zumbo's interface is the notch, expanding for history, settings and the Teach form and retracting again. MacWhisper is a window, which is the right shape for working through a file.
Account required
Neither asks you to sign up. Both are bought and activated, which is increasingly unusual and worth noticing.
Languages
You pick your main language during setup and Zumbo then writes only in that alphabet.
macOS minimum
Zumbo needs Apple silicon because the engine runs on the Neural Engine. MacWhisper does not state a minimum on the pages checked.
Download size
The engine ships inside the app rather than being fetched on first run, which is why the download is large and why it works with the network off.
Who should pick which
Pick MacWhisper if the thing you have is a file: a recorded interview, a call you captured, a podcast episode, a lecture. That is what it is built for, it does it locally, and there is a free version to try it with.
10:32Speaker 1We are holding the release until the migration lands.
10:33Speaker 2Then the changelog goes out Thursday, not Tuesday.
10:33Speaker 1Agreed. I will move the announcement.
Pick Zumbo if the thing you have is a cursor: a message to answer, a commit to describe, a document to draft, a meeting happening right now. Dictation is a different shape of tool and it is the one that changes a working day rather than a single task.
Owning both is reasonable and not expensive. They overlap almost nowhere.
Related: all six Mac dictation apps side by side, local meeting transcription, and Zumbo against Superwhisper.
Is MacWhisper a dictation app?
It is a transcription app. It writes out recordings you already have. Dictation, meaning words arriving where your cursor is while you talk, is a different job.
Can I use both?
Yes, and quite a few people should. MacWhisper for the interview or the recorded call, Zumbo for everything you write during the day.
Which one handles meetings?
Both, at different moments. Zumbo transcribes while the meeting runs, with speakers labelled as it goes. MacWhisper works from a recording afterwards with speaker recognition.
Are both one time purchases?
Yes. MacWhisper has a free version and Pro at 64 euro once. Zumbo is fifteen dollars once for one Mac.
Do either upload my audio?
Neither states that it does. Both run locally, which is the reason to pick either one over a cloud service.
Free for 3 days, everything unlocked. Then $15 once. macOS 14 or later, Apple silicon.