Feature · Moderation
AI flags the noise. Humans decide the edge cases.
Every incoming group message hits OpenAI’s omni-moderation and Perspective’s toxicity scorer. Your rules decide the action. The review queue catches the gray zone.
01
Rules engine, not spaghetti.
Trigger types: keyword, regex, AI classifier, behavioral, heuristic. Actions: delete, mute for 24h, ban, quarantine for review, notify admin. Compose as many rules as you need — each one is a workspace-scoped row in the database.
Moderation
01
02
Review queue for the gray zone.
Items above the review threshold but below the auto-action threshold land here. Approve, remove, or ban with one click. Every decision writes to the audit log with timestamp, actor, and reason.
- ·Approve / Remove / Ban
- ·One-click decisions
- ·Full audit trail — exportable CSV
- ·Realtime updates via Pusher
Moderation
02
03
Captcha on join (Telegram).
New joiners solve a button-tap challenge within 60 seconds or get removed. Eliminates 90%+ of bot spam before a single message is sent.
Moderation
03