Model parts with your hands and your voice. Pinch to grab a part, open your palm to orbit, speak the dimension you want, then export a printable STL.
Pinch to grab a part and move it through space; depth follows your hand as you reach toward the camera. Open your palm to orbit, hold it flat to pan, use both hands to scale and twist.
pinch → grabopen palm → orbittwo hands → scale
A local grammar handles the words designers use every day. When a phrase falls outside it, a model reads your live scene and replies with edits to the parts in front of you.
"add a red box""on top of the cylinder""rotate 90 degrees"
Parts stay parametric. Drag a slider or speak a number and the model updates live, with undo and redo throughout. Export the whole scene or a single part as a binary STL.
live slidersundo / redoexport STL
Moldr logs what it heard and what it changed, so the state of the model is never a guess. Here is a short session: a cylinder becomes a finished hub, by voice alone.
10:42:01youadd a cylinder
10:42:01moldrcylinder added, r 20, h 40
10:42:07youmake it sixty millimeters tall
10:42:07moldrheight set to 60 mm
10:42:14youbore a 12 mm hole through the center
10:42:15moldrhole cut, Ø12, through
10:42:22youpaint it slate blue
10:42:22moldrcolor set #5a86c2
10:42:29youexport
10:42:29moldrmoldr_scene.stl ready
Hands and voice drive everything. The mouse stays available as a full fallback, never required.
Say bigger, on top, or turn it. The grammar covers the words beginners and experts both reach for.
Hand tracking runs in your browser. The camera feed never leaves your computer; only fallback phrases reach the model.