Server Assistant🕵️ Alt / ban-evasion detection (altguard)
You ban someone, and ten minutes later a fresh account with the same avatar walks back in. Altguard is the bit that notices. When it’s switched on, every new member is quietly checked against the people your server has already kicked or banned — and if the match is confident, the alt is banned again before they can settle in.
It runs entirely on your own server’s moderation history. Altguard never reaches out to any shared list, central database or cross-server network — it only knows the people you have removed. That local-only design is the whole point of this feature, and we spell out exactly what it does and doesn’t see further down.
Who can switch it on: the server owner, or an admin with Manage Server. It’s off by default — altguard is opt-in, so nothing happens until you turn it on with
/altguard mode: on.
FREE available on every plan · ADMIN only an admin or the owner can toggle it
How the detection works
When someone joins, altguard compares them against each prior offender on file for your server and builds a match score out of 100. Several signals stack up — the more they line up, the higher the score:
- Reused custom avatar — the strongest signal. If the joiner’s custom avatar is byte-for-byte the same image as one a banned or kicked member used, that alone is a heavy hit. People rebrand their username far more readily than they bother to change their profile picture, so a reused avatar is the clearest “this is the same person” tell altguard has.
- Name similarity. A near-identical username or display name to a prior
offender counts strongly; a loosely similar one counts a little.
Toxic_Mikerejoining asToxicMike2won’t slip past on a name tweak alone. - Brand-new account. If the joining account was created within the last 7 days, that adds to the score — throwaway alts are usually freshly minted.
- Recent offence. If the original ban or kick was within the last month, the match weighs more heavily, since evasion tends to happen soon after the removal rather than months later.
A weak coincidence on one signal won’t trigger anything. It takes a genuinely convincing combination — and a reused avatar or a near-identical name on a fresh account is exactly the kind of pattern that pushes a match into confident territory.
Two outcomes: auto-ban or flag
- Confident match → auto-banned. When the evidence is strong — a reused custom avatar, or a near-identical name on a new account, with a high overall score — altguard bans the alt automatically and posts a notice to your log channel so your team has a record.
- Weaker match → flagged for review. When the signals are suggestive but not conclusive, altguard takes no action on its own. It simply flags the join to your staff with the reasons, and leaves the decision to you.
For altguard to carry out an auto-ban it needs the Ban Members permission and a role positioned above the member it’s removing. If it spots a confident match but can’t act, it tells you in the notice and asks you to review manually.
The /altguard command
One command, three modes — turn detection on, off, or check where it stands.
Switch repeat-offender detection on or off, or check its status. When it's on, new members are checked against your server's own banned and kicked members — confident alt matches are auto-banned and posted to your log channel, weaker matches are flagged for you to review. Only an admin or the owner can change it.
| Argument | Type | Notes |
|---|---|---|
mode required | choice | on = start checking new joins, off = stop, status = show the current state and how many offender fingerprints are on file. |
Checking the status
/altguard mode: status tells you whether detection is on and how many prior
offenders your server currently has on file — useful for confirming it’s armed
before you rely on it.
What a match looks like
When altguard catches something, it posts to your log channel (or your staff chat if you haven’t set a log channel). The notice shows the join, the match score, which prior offender it lines up with, the signals that fired, and what was done about it — red when the alt was auto-banned, amber when it’s only a flag for you to review.
A confident match — auto-banned
A weaker match — flagged for review
If you decide a flag is a genuine alt, removing them is a normal
/ban or
/kick away — and that
removal becomes part of the same local history altguard checks against next time.
Local to your server — and only your server
This is the most important thing to understand about altguard, so it’s worth being blunt:
Altguard only ever looks at the bans and kicks made on your own server. It fingerprints the members you have removed, and checks new joiners against that list and nothing else. It does not read, share, contribute to, or consult any cross-server list, central database or shared threat network. An alt who was banned on someone else’s server — but never on yours — is invisible to altguard.
A couple of specifics that follow from that:
- It builds up from the moment you turn it on. Altguard can only match people your server has already removed, so a brand-new server with no ban or kick history has nothing to compare against yet. The list grows naturally as you moderate.
- Soft-bans are deliberately excluded. A
/softbanis a ban-then-immediate-unban used purely to clear someone’s recent messages — it’s meant to let them rejoin. So soft-banned members are not recorded as offenders, and altguard won’t flag them when they come back.
Want protection that does span servers — catching known bad actors before they’ve ever offended on yours? That’s a different feature with a different, opt-in trust model: the cross-server ThreatNet. Altguard is its purely local counterpart; the two are independent and you can run either, both, or neither.
See also
- Moderation & safety — bans, kicks, soft-bans and the rest of the removal toolkit that feeds altguard’s local history
- ThreatNet — the cross-server counterpart to altguard’s local-only checks
- Back to the Wiki hub