CheckHacks vs Grace Anticheat

CheckHacks is a free, server-side Minecraft anticheat that detects client-side mods and hack clients using the Sign Translation Vulnerability. Players do not need to install a client mod: the plugin briefly places signs with special translation keys and checks the client’s response. Administrators can choose METEOR, TRANSLATE, or KEYBIND detection modes for each mod and add custom detections in checkhacks.yml. DoubleCheck repeats scans that return DETECTED or PROTECTED to help reduce false positives. CheckHacks can check players after Grim, Vulcan, Spartan, or Matrix flags them, and supports custom commands and detection alerts. A browser editor running on the game server lets administrators review scan history, look up players, and start checks. Access uses one-time tokens, while scan records are kept in a built-in SQLite database. Named server platforms include Bukkit, Folia, Paper, Purpur, and Spigot, with Minecraft versions 1.21 and later listed. The maker warns that client patches can temporarily block the detection method. The project is MIT-licensed, and donations are optional.

Grace Anticheat is a free, self-hosted anticheat for Java Minecraft servers. It targets closet cheating, ghost cheating and blatant hacking through rules-based detection and statistical analysis of player behavior. Servers can run a proxy arrangement with a Velocity plugin and Paper companion plugin, or choose a standalone Paper plugin without Velocity. The proxy setup requires Velocity 3.3.0 or later, Paper 1.19 or later and Java 21 or later; the standalone listing covers Minecraft Java 1.19.x, 1.20.x and 1.21.x. Grace can alert staff in-game and send formatted Discord webhook embeds. Configuration supports teleporting players back on hard alerts, kicking on action alerts and optional automatic bans, disabled by default. Staff commands expose engine status, evidence summaries, alert controls, violation history and configuration reloads. Grace is beta 0.1.0 and actively developed. Its API and related components are open source, while detection logic, model weights and threshold tuning are closed source.