Keep reading.
Just talk.

Free · Speech stays on your Mac · Apple silicon, macOS 14 or later
Try a character, then scroll
Where it stands, the words go.

One delivery

The moment you press, it knows where the words go.
While you talk, select what you mean.
Let go, and the words land where you were typing.
Enter is yours.

The details

While you talk,
it’s right there.

Covered up? It peeks out.
Talk first, pick the box after.
Click a window. It goes along.
A long talk? Tap, don’t hold.
Stop to think. It waits.
Didn’t catch it? It tells you why.
Said it wrong? Esc.
Slow to make out? It doesn't let go.

It listens

Characters

Pick the look. It means the same.

Click one, and the whole page wears it. The cat comes with the app, free for good.

The main window

Everything you said
is kept.

Day to day it stays out of the way: no Dock icon, and no menu bar icon unless you want one. What you said, the characters and the settings are all in the main window.

TodayHow much you said today, when, and to whom.
HistoryEvery line with the quotes and screenshots it carried. Search it, play it back. A line it couldn’t make out keeps its recording to try again.
Fix it onceIn History, select the misheard word and type it right once. From then on it is right.

Today

By hour
Just saidAll
To whom10 days

Today

Privacy

Wi‑Fi off. Still talking.

Speech is made out on this Mac. Recordings never leave it, and its logs never hold what you said. It goes online only to download the speech model and characters, to buy, and to check for updates once a day. Privacy policy

Almost ready.

Apple silicon (M1 or later) · macOS 14 or later · Hears Chinese, English and the two mixed · Notarized by Apple · The speech model, about 690 MB, downloads when you first open it

Once it’s out, this is how it goes, from download to your first line:

  1. 1Download
  2. 2Drag it to Applications
  3. 3Let it learn to listen
  4. 4Do what it asks
  5. 5Say your first line
Want a Windows version? Join the waitlist

Questions

What is digo?

digo is a free Mac app made for talking to AI agents: hold fn and talk, let go, and your words land where you were typing. While you talk, you can select text on screen to quote it, or click a window to send a screenshot along. Speech is recognized entirely on this Mac; neither the recording nor the text is uploaded. A small character stands where the words will go and moves as you talk; you can change how it looks.

Can I get it now?

Not yet. It is almost ready. Leave your email on this page and you will be the first to know.

Is the free version limited?

No. It is free to download and use, with no limits and no account.

Do the other characters cost money?

The cat comes with the app, free for good. Other characters are bought in the app (coming soon). They change the look and the voice, never what it does.

Does it rewrite what I say?

It does not rewrite, summarise or polish. It does a few small things: drops “um” and “uh”, spaces Chinese from English, writes github as GitHub. Each one can be turned off in Settings.

Which apps does it work in?

Anywhere you can type. Where it can read the caret, it stands by it; otherwise it stands on that window. Chrome and apps like VS Code too. Nowhere to type when you press? Click a box after, and it writes there.

Where is what I said kept?

On this Mac, in ~/Library/Application Support/Digo, with the recordings. Clear them in Settings any time.

I talked but nothing came out?

Check what it said: it tells you if the mic is muted, if you were too quiet, if it heard only noise, or if the mic switched. If recognition itself went wrong, the recording is kept so you can try again in History.

The words went somewhere else?

They go where you were typing when you pressed fn. Clicking elsewhere does not change that; if that place is gone, it copies the words for you.

Why Accessibility access?

To know you are holding fn, to find the text box and caret (that is where it stands), to read what you select (it becomes a quote), and to paste the words in.

What languages does it hear?

Chinese, English, and Chinese with English mixed in. The character speaks English or Chinese, following your system language; the app’s windows are in Chinese for now.

What recognizes the speech? Does it need the internet?

Qwen3-ASR 0.6B, a speech recognition model that runs on your Mac. It downloads once, about 690 MB, the first time you open digo; after that, recognition needs no connection.

Which Macs does it run on?

Macs with Apple silicon (M1 or later), on macOS 14 or later.

No fn key, or rather not use it?

Change it in Settings: right ⌥, ⌘, ⇧ or ⌃, F13 to F15, or just press the key you want.

Do I have to hold it for a long talk?

No. Tap it once and it keeps listening; tap again when you are done. Up to 10 minutes at a time. You can start the next line while the last is still being made out; they go in in order.

Who presses Enter?

You do. It writes into the box; sending is yours.