The best anticheatfor Minecraft.
DGuard learns from real cheaters: its combat model is trained on a live server with 500+ players online, and every player is watched on every front at once. Any new bypass is closed within a couple of days, and support answers you directly.
Free up to 30 players online · no card · set up in 2 minutes
DGuard panel example: a violation log with demo data.
Check by check
47 kinds of cheats — a check for every one.
We checked every kind of cheat DGuard has a check for against the source code and docs of GrimAC, Vulcan, Polar, Intave, Matrix. The same counting rules apply to everyone; sources are below the table.
Not a single miss.
- GrimAC22 / 47
- Vulcan25 / 47
- Polar18 / 47
- Intave27 / 47
- Matrix25 / 47
| Cheat · from public sources | DGuard47/47 | GrimAC22/47 | Vulcan25/47 | Polar18/47 | Intave27/47 | Matrix25/47 |
|---|---|---|---|---|---|---|
| Combat15 | ||||||
| KillAurahits targets by itself | Heuristics | partly or indirectly: no dedicated check — only indirectly, via Reach and packet order | ||||
| Multi-auraseveral targets in one tick | Heuristics | partly or indirectly: exists, but experimental — off by default | partly or indirectly: not named — only a generic cloud 'combat heuristics' | partly or indirectly: no dedicated check — only a general attack cap, never punished | ||
| TriggerBothits the instant the crosshair lands | Heuristics | partly or indirectly: not named — only the generic combat-automation heuristic | ||||
| AimAssist / Aimbotaims for the player | Heuristics | partly or indirectly: only two rotation artifacts: a 360° snap and a repeated look | partly or indirectly: not named — only a generic cloud 'combat heuristics' | |||
| Reachhits beyond reach | AttackRaytrace | |||||
| HitBoxenlarged hitboxes | AttackRaytrace | partly or indirectly: not named — only the hit-distance check | ||||
| AutoClickerclicks for the player | ClickPatterns | |||||
| Criticalscrits without a real jump | Physics · ProtocolScanner | partly or indirectly: no dedicated check — only the generic on-ground check | partly or indirectly: not named — only the ground-spoof check | partly or indirectly: no dedicated check — only the generic on-ground check | ||
| Velocityignores knockback | Physics | |||||
| KeepSprintsprint isn't reset after a hit | Heuristics | partly or indirectly: not named — only the generic movement simulation | partly or indirectly: no working dedicated check — only the generic movement simulation | |||
| SuperKnockbackextra knockback on the enemy | Heuristics | partly or indirectly: only the packet variant, via an experimental check | partly or indirectly: no dedicated check — only the generic sprint-packet checks | partly or indirectly: no working dedicated check | ||
| Backtrack / FakeLaghits past positions, holds packets back | AttackRaytrace | partly or indirectly: no dedicated check — partly, via the experimental BadPacketsR | ||||
| CrystalAura / AutoCrystalplaces and pops crystals by itself | 8 crystal rules | partly or indirectly: no dedicated check — only the hit distance to a crystal | ||||
| AutoTotema totem in hand in 0 ticks | InventoryClickAnalysis | partly or indirectly: no dedicated check — only inventory clicks while moving | partly or indirectly: a check was added in 2021; today's config doesn't have it | partly or indirectly: not named — only a generic inventory check | ||
| AutoPotion / AutoPearlthrows where the camera doesn't look | Heuristics | partly or indirectly: only a spoofed rotation inside the throw packet, from 1.21 | partly or indirectly: only how fast it throws, not where it looks | |||
| Movement13 | ||||||
| Speedfaster than the game allows | Physics | |||||
| Flyflies without elytra or creative | Physics | |||||
| NoFalltakes no fall damage | Physics | partly or indirectly: no dedicated check — only a spoofed on-ground flag | ||||
| Jesuswalks on water | Physics | |||||
| Stepsteps up full blocks | Physics | |||||
| Spiderclimbs walls | Physics | |||||
| Phasewalks through blocks | Physics | partly or indirectly: no dedicated check, by their own page | ||||
| NoSlowno slowdown when eating or blocking | Physics | partly or indirectly: only old sword-blocking, webs and honey — NoSlow while eating isn't detected | ||||
| ElytraFlycontrolled elytra flight | Physics | partly or indirectly: not named — the simulation recreates 'most' movements | ||||
| BoatFly / EntitySpeedflies and speeds up on vehicles | Physics | |||||
| Timerspeeds the game up | Timer | |||||
| Blinkholds packets, then teleports | Timer | partly or indirectly: no dedicated check — partly, via the experimental BadPacketsR | ||||
| OmniSprintsprints in any direction | Heuristics | partly or indirectly: not named — only the generic movement simulation | ||||
- GrimAC
Free and strong at movement. But it has no KillAura, Aim or AutoClicker checks — they're placeholders in its config, and heuristics are promised for a paid version that isn't on sale.
DGuard20 combat-behaviour rules plus aim, reaction and autoclicker checks out of the box — no experimental toggles.
- Vulcan
$19.99 per server. By its own page it targets “blatant cheaters who use common and publicly available hacked clients”; Phase, Blink, InventoryWalk and vanilla NoSlow are on its not-detected list.
DGuardPhase, Blink, InvMove and NoSlow are all checked, and auras with silent rotations get their own combat rules.
- Polar
Cloud-based: player packets go to Polar's servers. Movement and reach are never punished, there's no free tier, and it isn't sold in Russia or Belarus.
DGuardRuns entirely on your server, punishes movement and reach as well as combat — and is available in Russia and Belarus.
- Intave
Catches aura and scaffold only on a paid plan — €0.79 per 1,000 player-hours. Its machine learning runs in the cloud, also for a fee.
DGuardAura and scaffold are caught on every plan, and the combat model runs right on your server.
- Matrix
From $22, focused on 1.8.8 and 1.12.2. Its public checks list hasn't changed since 2022.
DGuardBuilt for 1.21.11 — the client taken apart down to the packet, with new rules coming in updates.
— a check in the source code or docs, — not in public sources, — partly or only indirectly, not counted (hover the icon for the reason). The rows are every kind of cheat DGuard has a check for; the same counting rules apply to everyone. Sources: GrimAC: checks, GrimAC: punishments.yml, Vulcan: resource page, Vulcan: check descriptions, Polar: detections, Polar: Russia and Belarus, Intave: checks, Intave: pricing, Matrix: docs, Matrix: checks.yml.
The cost of waiting
Without DGuard, you lose every day. Even if you can't see it yet.
A cheater never leaves alone: honest players go with them, and so do their purchases and your server's reputation. A weak anticheat isn't free — you pay for it with the players it let down.
Players
Without DGuard: An honest player gets killed by KillAura twice — and quietly moves to another server. They won't file a report.
With DGuard: The aura gets banned before the player even realises they were cheated. They stay and keep playing.
Revenue
Without DGuard: Your server runs on honest players — exactly the ones cheaters drive away. Their purchases leave with them.
With DGuard: Honest players stay — and keep supporting the server.
Reputation
Without DGuard: “The server is full of cheaters” — one review like that on a server list scares off newcomers more than any ad can win them.
With DGuard: Newcomers see fair play from their first join — and stay.
Staff time
Without DGuard: Moderators spend hours watching replays and arguing with players instead of building the server.
With DGuard: Every ban comes with a report that explains the evidence. An appeal takes minutes.
Economy and leaderboards
Without DGuard: Nuker, FastBreak and AutoClicker inflate resources and leaderboards — honest players stop trying.
With DGuard: Block-breaking, click and movement checks keep the economy fair.
Winning a player back is harder than keeping one.
Protect your server freeWhy DGuard
6 reasons to switch to DGuard.
What DGuard does differently — and why cheaters don't last with it.
Evidence from every front
Combat, movement, packets, inventory, building, timing: 11 checks and 80+ rules watch every player at once. Hiding one cheat won't save them.
Aggressive — and invisible
No cancelled hits, no setbacks. The cheater has no idea they're caught while DGuard keeps stacking evidence.
No snap judgments
Every flag is evidence with a weight: VL × the rule's weight, halving every day. A ban comes only at 400 points across three separate episodes. One dossier across all your servers.
- ban threshold: 400
- KillAuraBan
- SpeedKick
- AutoClickerFlag
Precise punishments. Your rules
Ban or kick per cheat feature, evidence weights and thresholds — all in the config. From lenient to zero tolerance. Try switching.
Learns every day
Flag logs and fight recordings from a live server turn into new rules every day, and thresholds are tuned on thousands of players. Cheaters stand out — honest players sleep easy.
Support without middlemen
You write straight to the DGuard team — in the panel, on Telegram or via the website form. A new bypass is taken apart from fight recordings and closed with a rule in an update.
Packet order
Cheats give themselves away by packet order.
We took the Minecraft 1.21.11 client apart down to the packet. A legit client sends packets in a strict order. A cheat breaks it — and gets caught, no matter how human it pretends to be. Lag can't do that.
- 1Attack
- 2Rotation
- 3Movement
- 1Rotation
- 2Attack
- 3Movement
combat rules DGuard adds on top of the base checks
- Rotation packet before the hitKillAura with silent rotations
A client attacks at the start of a tick and sends its rotation after. The reverse is impossible.
- Target trackingAimAssist
A human hand catches up with the target in 3–5 ticks. An assist does it in the same tick.
- Mouse step in combatAimAssist, aura
A mouse rotates in multiples of the sensitivity step. A cheat doesn't.
- Snap-back and 360° resetKillAura
Turning to the target and back within one tick: the angle “jumps.”
- Two targets per tickMulti-aura
A legit client picks one target per tick.
- Reaction to aimTriggerBot
The hit lands on the tick the crosshair does; 28 of 30 intervals match to the tick.
- Crit on the first possible tickTriggerBot with crits
Humans pull that off on about 21% of jump hits. A bot — on almost all of them.
- Crystal aura: 8 rulesCrystalAura, AutoCrystal, AutoTotem
Hitting a crystal before it appeared on the client, guessing entity IDs, hits past the edge of reach.
- Throws outside the line of sightAutoPotion, AutoPearl
The throw packet carries a rotation the camera never had.
Results
A week on live servers. Numbers straight from the logs.
What DGuard saw in a week of beta: how many players it checked, whom it punished and how precise its rules are.
- players in 24 hours
- 3,596
- SURVIVAL, Oct 12–13
- DGuard bans in the same 24 hours
- 322
- on the same server
- combat model hit rate
- 92%
- 11 of 12 players scored ≥ 0.95 were later banned independently
- of one core on the main thread
- 0.03%
- 0.38 ms of a 10.9 ms tick at 121 players
- Rejects server-set rotation86%
- Rotation packet before the hit77%
- Tracks target with no reaction delay64%
- Non-mouse rotation in combat60%
- Combat model58%
- Crit on the first possible tick55%
- Hits beyond reach36%
Dashed line — server average: 5.5%. Only players DGuard hadn't banned itself yet.
Beta logs from the SURVIVAL and ANARCHY servers, Oct 7–13, 2026.
Pricing
Full protection on every plan. Pay only for players online.
Free up to 30 players online, forever. Beyond that you pay for players online and the number of servers, not for protection.
Prices in US dollars. No auto-renewal, switch plans anytime.
- StarterFor a new server$0free, with no time limit
- up to 30 players online
- 1 server
- Full protection
- Ticket support
- Best valueServerFor a growing server$24/ moor $18 / mo billed for 3 months
- up to 150 players online
- 1 server
- Full protection
- Priority support
- NetworkFor a server network$59/ moor $44 / mo billed for 3 months
- up to 500 players online
- up to 5 servers
- Full protection
- Shared player dossier across servers
- Priority support
- Setup assistance
- EnterpriseFor top projectsCustomtailored terms
- 500+ players online
- unlimited servers
- Full protection
- Shared player dossier
- Dedicated support
- Tailored configuration
FAQ
Frequently asked questions
Versions, performance, bans, configuration and licensing — short and to the point.
Didn't find your answer?
Write to support — we'll look into your server and help with the setup.
Contact supportStop losing players. Protect your server today.
Cheaters won't wait while you pick an anticheat.