The AI Chat Filter reads context instead of matching a wordlist. It is decisive about scams and phishing while staying relaxed about casual banter.
Activate the filter, pick a mode and set the log channel.
What happens at mild, moderate and severe, and for repeat offenders.
Text, links and images: what a catch looks like.
Setup, per-user stats, config, clearing warnings and opting out.
Filter not acting, staff exemptions, logs and large servers.
A screenshot and a plain-English explanation for every option.