SayHi

Inspiration

Every interface we use is assembled from primitives — keys, clicks, taps. We wanted to add a set our computers have always ignored: the shapes our hands already make.

Hands are the oldest interface humans have. We gesture before we can speak, and we keep doing it long after — pointing, waving, giving a thumbs up. But the moment we sit down at a computer, all of that is thrown away, and control narrows to a keyboard and a mouse within arm's reach.

So we built the blocks. Thirteen handshapes, tracked separately for each hand, each one an empty slot the user fills with meaning.

What it does

SayHi watches your computers camera and turns hand gestures into actions.

It detects your hand, classifies the shape it's making, and fires whatever you've bound to it — take a screenshot, launch an app, hit play/pause, minimise a window, open a URL, send a keyboard shortcut. Because left and right hands are tracked as separate bindings, thirteen shapes give you twenty-six distinct slots: a left fist and a right fist can do completely different things.

Gestures also compose. Assign a macro and you get an ordered list of steps — open an app, wait for it, type a block of text, press ⌘Return — all behind a single gesture. Blocks made of blocks.

Everything runs on-device through Apple's Vision framework. Frames are analysed and immediately discarded. There is no network code in the app at all.