KillAura gives itself away.
GuardAC watches how a player moves the crosshair in a fight and tells a real hand from an assisted one. Players install nothing: the work happens in our cloud.
See it running on a real server
36 seconds on a live server: the detection during a fight, the staff alerts, the punishment and the dashboard.
Try it yourself
Move your mouse in the field below and watch GuardAC read the movement of your hand.
Try to move perfectly straight at one steady speed. A hand cannot do that, and the bar slides into the red.
Why it works
A cheat can put the crosshair on target, but it cannot move a mouse the way a hand does.
Catches KillAura by movement
Aim that lands too smoothly never looks human: a real hand shakes a little and overshoots the perfect line.
Everyone aims differently
The plugin learns how a player normally moves the mouse. When that style changes mid-session, it shows immediately.
Careful with honest players
Punishment only fires once the model is sure. Until then your staff get alerts and make the call themselves.
How it works
Three steps from install to the first alert.
The plugin watches the fight
While a fight is on, the plugin records how the crosshair moves. That is a short list of numbers rather than video, so the server barely notices it.
The neural net answers
Those numbers go to the GuardAC cloud, which returns the probability that a cheat is helping the player.
Your call
Staff get an alert with the name and the score. Keep the punishment in your hands or hand it to the automation, from a kick to a ban with an animation.
Pricing
Pick the plan that matches your server's average online. Start free and switch whenever you like.
FAQ
No. Everything runs server-side, and players never notice the anti-cheat is there.
Barely. The plugin sends a short list of numbers about crosshair movement, and the neural net itself runs in the cloud.
Yes. A key is issued right after sign-up, no card needed.
Paper, Spigot, Purpur and Folia from 1.16 up, on Java 17 or newer.
Pick a plan and a term (1, 3, 6 or 12 months) and pay by SBP or crypto. Promo codes work too.
The neural net looks at the physics of mouse movement in combat: how the crosshair accelerates, brakes and trembles. KillAura draws those curves too cleanly, and the model sees the difference within seconds.
The model judges a run of combat windows instead of one moment, and acts only when several windows in a row agree. Thresholds are yours to set, from alert-only to strict auto-bans.
Run the alert-only mode and punish by hand. Or let the automation handle it and write your own ladder of punishments in punishments.yml.