Say it. It's done.

SayIt turns speech into clean text in any app and runs a voice-controlled AI agent on your Mac. Hold the fn key to dictate; double-tap it to give the agent a task.

Any Mac running macOS 13 Ventura or newer, Apple Silicon or Intel (one universal download). A Windows version is in the works.

Speech recognition covers more than 90 languages, and the cleanup keeps your language's spelling and punctuation. Mixing two languages in one sentence works, and you can add your own names and terms to the built-in dictionary.

It sees the app you're in, the text field, and (if you allow Screen Recording) the screen. It can type a reply, rewrite what you just said, remember your preferences per app, fill in forms, turn a blunt instruction into a polite customer reply, open apps and websites, click through interfaces, and run read-only commands. Anything that sends, submits, deletes, pays or changes files stops and asks you first: tap fn to approve, Esc to stop.

Passwords, API keys and tokens are masked on your computer before any text leaves it. Password fields are never read, and password managers are never captured. Your dictation history lives in your account and you can delete it from the app at any time.

Microphone and Accessibility are required. Screen Recording is optional: without it the Agent works from the text field and the accessibility tree; with it, the Agent looks at the screen only when a task needs it.

Free: 200 actions to try, with no time limit. When you reach the limit, the app shows the Premium plans and you can upgrade right there by card or PayPal (bank transfer in Vietnam). One action is one dictation or one task for the Agent.