started building Precis after using apps like Wispr flow and realizing how much I liked the idea of voice being a real input method, not just transcription.

Then Apple opened up its on-device Foundation Models, and that got me thinking: if the intelligence can run locally, why should a simple everyday tool need a cloud AI bill?

So I built Precis around that idea. The speech recognition and rewriting happen on the Mac, using Apple’s on-device models where available. No API costs, no subscription for AI usage, and it keeps working without an internet connection.

I also wanted something I’d actually use every day, so the goal became pretty simple: hold a shortcut, talk normally, release, and keep working.

I’m building this as a solo developer, so I’m particularly interested in where it falls short. Human beings apparently enjoy finding bugs in other people’s software. This is your opportunity.

started building Precis after using apps like Wispr flow and realizing how much I liked the idea of voice being a real input method, not just transcription.

Then Apple opened up its on-device Foundation Models, and that got me thinking: if the intelligence can run locally, why should a simple everyday tool need a cloud AI bill?

I also wanted something I’d actually use every day, so the goal became pretty simple: hold a shortcut, talk normally, release, and keep working.

I’m building this as a solo developer, so I’m particularly interested in where it falls short. Human beings apparently enjoy finding bugs in other people’s software. This is your opportunity.

Precis for Mac

Mac dictation that writes for the app you're in

Mac dictation that writes for the app you're in