Docs · Modules

Moderation & AutoMod

Warn, timeout, kick and ban with case numbers and a mod log – plus AutoMod against spam, links and more.

Moderation

All actions via slash command or in the dashboard (Mod cases, Warnings). Every action gets a case number and lands in the mod log channel (Dashboard → Settings).

  • Automatic actions: after a certain number of active warnings, a timeout, kick, temporary or permanent ban follows automatically (Dashboard → Warnings → Thresholds).
  • Temporary bans lift themselves – even after a restart.
/warn <user> <reason>Mod
Warn a user
/infractions <user>Mod
Show a user's warnings
/timeout <user> <duration> [reason]Mod
Time out a user
/kick <user> [reason]Mod
Kick a user
/tempban <user> <duration> [reason]Mod
Ban a user temporarily
/ban <user> [reason] [delete_days]Mod
Ban a user
/unban <id> [reason]Mod
Unban a user
/clear [amount] [user] [all]Mod
Delete messages

AutoMod

Dashboard → AutoMod. Turn rules on individually and adjust them:

  • Anti-spam (messages per time span), mass mentions, foreign invites, links, capital letters, word filter
  • Per rule: delete the message, warn or timeout
  • Exempt roles and channels; members with “Manage Messages” and admins are always exempt

Common problems

“Missing permission”. The ZoneBuddy role must be above the member’s role.

AI moderation

In addition to the word filter, AutoMod can have messages checked by a moderation AI – it detects insults, hate, threats and the like even without a word list and in paraphrases.

Dashboard → AutoMod → AI moderation:

  • Categories: insults & harassment, hate & incitement, violence & threats, sexual content, illegal content, self-harm. Sexualisation of minors is always detected.
  • Sensitivity: low (only clear cases), medium or high (detects more, but sometimes catches jokes).
  • Action: report to the team only (message stays, notice with a link in the report or mod log channel), delete, delete + warn or delete + timeout.
  • Self-harm is never punished – the team only gets a notice so they can respond with empathy.

Important: With AI moderation on, the message text is sent to OpenAI (only for checking, not for training). Moderators and exempt roles and channels are not checked.

Scam image block

The AutoMod rule Scam image block detects known scam images (fake Nitro, “free giveaway”, crypto scams) – even if they were slightly altered, resized or re-saved. ZoneBuddy compares a fingerprint of the image (perceptual hash); only this fingerprint is stored, never the image.

  • Block an image: right-click a message → Apps → Mark as scam (requires “Manage Messages”). The message is removed and the image lands on your server’s block list.
  • Waves: if several new accounts (default: 3 accounts in 10 minutes, account younger than 30 days or on the server for less than a week) post the same image, ZoneBuddy blocks it automatically and cleans up all messages of the wave.
  • Shared list: if three different servers have blocked the same image, it automatically protects all servers that use the shared list (default: on).
  • Matches are deleted, depending on the settings with a warning or timeout, and reported to the team (report channel or mod log). In the dashboard you see your server’s blocked images and can unblock them.

Tip: Single-color images and simple gradients are never blocked – they would otherwise match each other.