Before the
command.

We build the layer that connects a gesture, a few words, and the action a person means to take.

Explore our research

A fingertip gives
“this” a meaning.

When someone points and says “move this,” the gesture identifies the file. A second gesture can give it a destination. We build human-to-machine intent infrastructure so those cues can guide an action across devices, operating systems and AI models.

Continuous intent

Align a word with the gesture that accompanies it and the object on screen at that moment.

Incremental action

Prepare a reversible preview while the destination is still taking shape. Ask for confirmation when it matters.

Continuous correction

Let a second gesture change the destination while keeping the original task in view.

“Move this over there.”

The same request can take two paths. Watch a file move, or change its destination before the move is confirmed.

Hearing the request

The word “this” leaves the file unspecified. The system waits for a pointing gesture.

File identified

The fingertip and the spoken word refer to the same file. The destination is still open.

Waiting for confirmation

A second gesture selects a folder. Choose either destination to revise the preview.

Move verified

The workspace confirms the file arrived in the selected folder.

Following a correction

The file stays selected. A new gesture asks the system to look for a different folder.

Workspace“Move this…”
Field recording.wav

Across devices and software

Our SDK and API research explores how one intent model can serve different applications, agents and devices. Each integration checks permissions, routes the action and reports what actually happened.

Close-up of the graphite Syno Ring R1 worn on a finger.

Syno Ring R1

A small movement of the thumb becomes an input for AR and smart glasses. Syno Ring brings our work on gesture into a device worn on the finger.

Explore Syno Ring