A motorized mousepad that auto-aims with computer vision

The creator of the YouTube channel Basically Homeless showcased an unusual aim-assist system that quite literally moves the mouse by itself—thanks to a motorized mousepad. Instead of meddling with the game or the mouse firmware, blogger Nick leaned on computer vision and a mechanical platform that physically slides the pad under the mouse, snapping the crosshair into place with surgical precision.

The setup analyzes the on-screen image using a YOLO-based computer vision model, detects targets, and calculates how far and in which direction to shift the aim. The software then sends commands to a motion controller, and the pad jumps into position almost instantly, turning aim corrections into picture-perfect headshots. It looks so natural that opponents were convinced they had suddenly started playing at an esports level.

At the heart of the build is a repurposed Xdraw A4 drafting tablet paired with a GRBL controller, a combination more common in CNC devices. Nick cut a hole in the desk, installed the motion mechanism underneath, and mounted a glass mousepad on top. He used 3D-printed brackets for fine alignment; otherwise, the rig proved surprisingly straightforward and reliable.

The most telling moment came in a test with a friend named Olivia. While the assist was active, she was certain her in-game prowess was entirely natural; once it was disabled, her results fell off a cliff, which immediately prompted suspicions of cheating—after that, the assist went back on, ostensibly in the name of fair play.

The project is a clear illustration of how a blend of mechanics, software, and computer vision can sidestep traditional anti-cheat tools and deliver an almost invisible yet remarkably effective form of auto-aim. Ingenious—and just unsettling enough to make competitive players take notice.