AI text moderation
Understands context, so it catches what keyword filters miss.
- Spots subtle harassment, threats and coordinated spam, not just keywords
- Categories you configure, each with its own confidence threshold
- Three actions: flag for staff, delete or mute
- Its reasoning goes in the mod log, with one-click Warn, Timeout and Ban buttons
- Immune roles, and channel scoping by blacklist or whitelist
- Reads the conversation around a message to cut false positives