Every server owner eventually hits the same moment: someone is flying around spawn, or winning fights they shouldn't, and the regulars are asking what you're going to do about it. The instinct is to install whatever tops a "best anti-cheat" list. That can work, but the anti-cheats that suit a 20-player survival server and a 300-player PvP network are rarely the same. A few questions up front save a lot of pain later.
Start with your server, not the feature list
Before comparing anti-cheats, write down the facts about your own setup:
- Server software and version: Paper, Spigot, Purpur, or Folia, and which Minecraft version. Many anti-cheats support some of these and not others, and Folia support in particular is uncommon.
- Client versions: if you use ViaVersion, ViaBackwards, or ViaRewind to let older or newer clients join, check that the anti-cheat supports that range too, not just your server's version.
- One server or a network: a BungeeCord or Velocity network needs an anti-cheat, and a ban system, that work across servers.
- Bedrock players: if you run Geyser and Floodgate, find out how the anti-cheat treats Bedrock players. Many simply ignore them (more on that in our Geyser guide).
- Game mode: competitive PvP needs strong combat checks; a survival or creative server mostly needs movement checks and a light touch.
That list alone usually rules out half the options.
False positives matter more than the number of checks
An anti-cheat that misses a cheater is annoying. An anti-cheat that bans a loyal regular is worse: that player tells their friends, and your community starts to distrust your moderation. When you compare options, ask how each one handles the things that make honest players look suspicious: high ping, server lag spikes, knockback from other plugins, and translated clients. We wrote a separate guide on why anti-cheats flag honest players, and it's worth reading before you turn on automatic bans with any tool.
Look for evidence your staff can act on
A notification that says "Player failed Reach (VL 12)" tells a moderator very little. Better anti-cheats show what the player actually did, when, and why it was flagged, so a staff member can judge it and explain the decision if the player appeals. If your team is small, this matters even more: you want alerts you can trust at a glance, not a stream of numbers to decode.
Check the performance cost
An anti-cheat runs for every player, every tick. A heavy one can cost you more lag than the cheaters did. Look for published performance figures, prefer anti-cheats that do their work off the main server thread, and measure it yourself with a profiler such as spark before and after installing.
Maintenance matters as much as detection
Minecraft updates often, and every update can break an anti-cheat or open new bypasses. Check how quickly the anti-cheat supported the last few versions, when it was last updated, and whether there's an active support channel. An anti-cheat that was excellent two years ago and hasn't been updated since is a risk.
Free or paid?
Price isn't a reliable signal of quality in either direction. There are capable free, open-source anti-cheats, and well-supported paid ones. What you're really comparing is support, update speed, and how much tuning you're willing to do yourself. Whatever you pick, buy or download it from the developer's official page or a reputable marketplace, never a reupload.
A simple way to shortlist
- Pick two or three anti-cheats that pass the compatibility questions above.
- Install one at a time on a test server, or on your live server with punishments switched off and alerts only.
- Play normally for a week, including on a bad connection and as a Bedrock player if you have them, and watch what gets flagged.
- Keep the one whose alerts your staff actually trust, then turn on punishments gradually.
Where Argus fits
We're building Argus, a server-side anti-cheat for Paper servers, around the priorities above: evidence behind every finding, a light footprint, and fair handling of Bedrock players. It's honest to say where it stands: v1.0 is in development and hasn't been tested with real players yet, so it isn't something to install today. If it sounds like a fit, you can follow along or pre-order on the Argus page. Whatever you choose, the shortlist process above will serve you better than any single "best of" list.