AI moderation

Off by default

An AI toxicity filter for automod: Llama Guard checks messages for hate, threats and worse.

On this page

Set it upLink to this section

Open the dashboard and pick your server. Sign in with Discord if you haven’t.
Find AI moderation in the module list (press / to filter it, or ⌘K and type its name) and turn it on.
Go through the settings below. Simple mode shows the essentials; switch to Advanced for the rest.
Changes apply in your server within a few seconds of saving.

What you’ll find on the AI moderation page of the dashboard. Settings marked Advanced only show in Advanced mode.

CategoriesList: What Kovu acts on. S1: Violent crimes, S2: Non-violent crimes, S3: Sex-related crimes, S4: Child sexual exploitation, S5: Defamation, S6: Specialized advice, S7: Privacy, S8: Intellectual property, S9: Indiscriminate weapons, S10: Hate, S11: Suicide and self-harm, S12: Sexual content, S13: Elections, S14: Code interpreter abuse.
Log channelChannel

Delete flagged messagesOn/off
Warn in the channelOn/off: A short notice that deletes itself.
Give a formal warningAdvancedOn/off: Creates a warning case, which counts toward autopunish.
Flags before punishmentNumber: Flagged messages from one member within the window.
WithinDuration
PunishmentChoice
Mute lengthAdvancedDuration: 0 uses the default from Moderation settings.

Ignored channelsChannels
Ignored rolesRoles
Only these channelsAdvancedChannels: Leave empty to check every channel.
Only members who joined withinAdvancedDuration: Leave at 0 to check everyone.

Keeps AI checks fast and within fair use.

Messages checked per dayNumber: Anything past this in 24 hours goes unchecked by AI (other filters still run).
Shortest message checkedNumber: In characters, not counting links, mentions and emoji.
Check each member at most everyDuration
Share of messages checked (%)Number

Pick your server, turn it on and save. Changes reach Discord in seconds.