Skip to main content
Guard is the moderation bot. Everything you need to run a tight server with a small team.

​
Core moderation

All actions are source-agnostic — whether you trigger them from the dashboard, the /mod slash command, or AutoMod, they produce the same case row with the same case number.
ActionWhat it does
BanPermanent or time-bound. Auto-unban when the duration expires.
SoftbanBan + immediate unban to purge the user’s recent messages.
KickRemoves the member. They can rejoin with a new invite.
MuteDiscord timeout. 60s to 28d.
WarnLogs + DMs the target. Warns decay after N days (configurable).
NoteStaff-only note. No DM, no Discord-side action.
Purge / Lock / Unlock / SlowmodeChannel-level ops, audit-only.
Every action writes to mod_cases with a monotonic case number per guild. The Case Browser lets you filter by action, source, moderator, or user.

​
AutoMod

Ten rule types, all with shared config surface (triggers, exempt roles, exempt channels, action chain):
  • Spam — rate-limit messages per user.
  • Mention spam — cap unique mentions per message.
  • Links — block all, or allow-list specific domains.
  • Invite — block Discord invite links with optional whitelist.
  • Regex — custom pattern matching with tested preview.
  • New account — flag users whose account age is below a threshold.
  • Image only — restrict a channel to image-only messages.
  • Caps — flag messages with too much uppercase.
  • Emoji spam — cap repeated emoji per message.
  • Duplicate messages — detect copy-paste across channels.
Actions chain: delete → warn → mute (with duration) → kick → ban. Every trigger surfaces in the Triggers feed with the message content, channel, and action taken.

​
Anti-raid

Lockdown mode. One click from the dashboard or /raid lock slash command:
  • Raises the server verification level to Highest.
  • Records the previous setting so unlock restores it.
  • Announces the lockdown with reason + expected duration.
Mass ban. Paste up to 500 user IDs, optional reason + message-delete window. All bans run through the Action Queue with per-moderator rate limits respected.

​
Anti-nuke

Protects against compromised admin accounts. Quotas per event type with audit-log actor lookup:
  • Channel deletes
  • Role deletes
  • Bans
  • Kicks
  • Webhook creates
When a user exceeds the quota within the window, Guard:
  1. Strips their roles, bans them, or moves them to a quarantine role (your choice).
  2. Optionally auto-reverts ban/kick actions they took while over quota.
  3. Posts an alert to the configured channel with actor, event count, and revert summary.
Whitelist roles, whitelist users, and ignore-bots are all supported.

​
Verification gate

Post a button message. When new members click:
  • Account age is checked against your minimum (e.g. 7 days).
  • If pass, the configured verified role is assigned.
  • If fail, they get an ephemeral message explaining why.
Combined with Discord’s member-screening, this kills drive-by spam accounts without any friction for real members.

​
Logging

By default, all moderation events go to one channel. Premium upgrade: per-event routing. Send:
  • Mod actions → #mod-log
  • Member joins/leaves → #join-log
  • Channel/role/emoji CRUD → #audit-log
  • AutoMod triggers → #automod
  • Voice events → #voice-log
  • Anti-nuke alerts → #alerts
Every embed carries the actor’s avatar, the action, the target, and a clickable case number where applicable.

​
Welcome + leave

Pick a channel, write a message (tokens like {user}, {server}, {server.membercount} supported), optionally post an embed. Auto-role on join applies a role to every new member after verification.

​
Custom commands

Text or embed responses, triggered by /<name> slash commands. Token language:
  • {user} — mention the invoker
  • {server} — server name
  • {channel} — current channel
  • {random:A|B|C} — pick one at random
  • {if_role_has:ID:then|else} — conditional on role
Allowed-role and allowed-channel restrictions per command. Cooldowns per user.