Talk, and it types. Anywhere on your Mac.
Hold one key, say what you mean — however messily — and clean written text lands in whatever app your cursor is already in.
It sits at the edge of the screen. Out of the way.
No window to open, no app to switch to. Sotto waits where you're already working.
Hold one key and start talking.
An email, a chat, a code comment, a form field. It listens wherever your cursor is.
Say it like a person.
um so can you send the invoice to Mark uh tomorrow morning no wait, before ten
It arrives written.
Can you send the invoice to Mark tomorrow, before ten?
Not a transcript. A finished sentence.
Plain speech-to-text hands your stumbling back to you. Sotto reads the intent and writes it the way you would have, given the time.
Hold the key, pick a number. Same words, shaped for where they're going.
Hover a key to preview, press to play the full demo with voice.
The rest of what Sotto does
Talk to your calendar
Say it out loud — Sotto reads the day, the time and the intent, and writes it straight into Reminders or Calendar.
Any language in, any out
Speak one language, type another — translated on the way to the page.
Examples: spoken Vietnamese "Gửi báo giá cho khách hàng" is typed as English "Send the quote to the customer." Spoken Polish "Potwierdzam dostawę w poniedziałek" is typed as Chinese "确认周一交货。" Spoken English "Running five minutes late" is typed as German "Ich bin fünf Minuten zu spät."
Read text off the screen
Drag a box over anything you can see — a locked PDF, an image, a slide, a field you can't select — and Sotto lifts the text out and cleans it, all on your Mac.
A clipboard that stacks
Copy several fields in a row, then paste them back in order.
Three fields are copied one after another onto a stack, then pasted into a new form in the order they were copied.
Message your team, hands-free
Say who it's for and what you mean. Sotto writes the message, puts it where it belongs — a mail draft, a chat box or a calendar invite — ready for you to send, without leaving what you were doing.
It learns how you talk
Your brands, your people, your warehouse words. Say one three times and Sotto keeps its spelling for good — and shows you exactly what it picked up, so you can make it forget.
It knows where you're typing
Sotto reads the app you're in and the window's title — so Gmail sitting inside Chrome is treated as mail, not as a browser. Same thought, dressed for the room it lands in.
Select, tap, translated
Highlight anything on screen and press ⌥T. A sticky note opens where your cursor is, the original beside the translation, with a Copy button. Nothing is pasted over your work.
Your words stay yours
Every take is kept on your own disk, readable by your account and nothing else on the machine.
Notes for calls and meetings
Someone calls. Sotto is already in it.
No bot to invite, no link to paste. Pick up, and it is recording both sides from the first word.
Your voice and theirs, kept apart.
Your microphone is one channel, the call audio is another. Who said what is never a guess.
It writes while you talk.
Every line attributed, landing in the app before the call is over.
Talk over each other. It still knows.
One mixed recording blurs two people into one voice. Two channels keep both lines whole — even when they land on top of each other.
The call ends. The notes are written.
A summary, the decisions, and the follow-ups — and the one that was given a time goes straight into your calendar.
Example of a two-channel call transcript. An incoming call is answered. You: Can we ship the pallet before Friday? Them: Friday’s tight, Monday is safe. Then both speakers at once — You: Monday works if the invoice lands tonight; Them: I’ll send it within the hour. Summary: shipment moves to Monday, Friday not feasible; invoice to arrive tonight, within the hour. Action items: invoice to arrive tonight, on them; confirm the pallet left, on you, scheduled for Monday 18 August at 09:00.
One app, both moods. Pick one and the page follows.
Sotto has a light face and a dark one, and it wears whichever your Mac is wearing. Try it here — choose a side and this whole page changes with it. We'll remember which one you picked.
Audio goes one place. Only to be written down.
Your recording goes to the transcription service that turns it into text, and nowhere else. No analytics, no tracking, no quiet copy kept for training. History and settings stay on your disk, locked to your account. Crash reports are off unless you turn them on — and then they carry a stack trace, never a word you said.
Getting in
We're letting people in a handful at a time so we can actually answer you. Leave your email and you'll get a build and a direct line to us.
The full app, signed and notarized. Nothing to configure beyond the key you hold.
Transcription included — no keys to manage, no per-word bills. Cancel whenever.
For teams over 20, or anything bespoke. Tell us what you need and we'll shape a plan with you.
Talk to us