Voice typing for Gmail on Mac.
The reply you have been avoiding is usually four sentences. It is not hard to write, it is just slow to type, and there are eleven of them.
Why email in particular
Email is the purest transcription task in a working day. By the time you open the reply you already know what it says; you have known since you read the original. Everything after that is your hands catching up with a decision you have already made.
It is also the place where being slow makes the writing worse rather than just later. The reply that should have three sentences of context gets one, because the third sentence costs forty seconds you do not feel like spending, and then somebody asks the obvious follow up question and the thread is four messages long instead of one.
How it works
Hold it to talk, tap it to toggle, Escape cancels. Pick the key in Settings.
Gmail has no dictation of its own in the browser. Google Docs does, Gmail does not. System wide dictation covers it, and the same setup covers Mail and Outlook without any further work.
- Click into the body, the subject line or the search box.
- Hold your shortcut, which you chose in Settings during setup.
- Talk, including the parts you would normally shorten. There is no session limit.
- Let go. The text lands in about a fifth of a second, punctuated.
- Read it before you send, the same as anything else.
Read and talk at the same time
This is the trick that makes dictated email better rather than merely faster. Keep the original message on screen, look at it, and answer point by point out loud. You cannot do that while typing, because typing requires the part of your attention that reading uses. The replies that come out of this are more complete, not less, because you are answering what was actually asked.
Names, which is where email goes wrong
Email is full of people. A general speech model spells a surname as the nearest ordinary word, and sending a client their own name misspelled is worse than sending nothing. Teach fixes each one permanently in about three seconds: select it, type what you meant, press Teach. The corrections stay on your Mac and build up into the set of names you actually write to.
The quiet one: nothing is uploaded
Email contains other people's information by definition. Zumbo has no server for your voice: no account, no API key and no network request needed to transcribe, with the speech engine inside the app. The finished text is typed into the compose window and the audio never existed anywhere but your Mac.
Related: dictating into Google Docs, adding custom words to Mac dictation, and dictation with the Wi-Fi off.
Does Gmail have dictation on a Mac?
Not of its own in the browser. Google Docs has voice typing, Gmail does not, so the answer for email is dictation at the system level, which also means it works in Mail and in Outlook without changing anything.
Does it work in the subject line and in search?
Yes. Zumbo types where the cursor is, so the body, the subject, the recipient field and the search box behave the same.
Will it get names right?
Correct one with Teach and it is heard your way from then on. Colleague and client names are the most common thing worth teaching, and the corrections stay on your Mac.
Can I dictate a reply while reading the original?
Yes, and it is one of the better reasons to use it. You can read the message on screen and talk the reply, which is very hard to do while typing.
Does my email go anywhere before Gmail?
No. The speech engine ships inside the app and composes on your Mac. Zumbo types the finished text into the compose window.
Free for 3 days, everything unlocked. Then $15 once. macOS 14 or later, Apple silicon.