Write moderation rules the way you would brief a person
Describe what should be hidden, flagged, answered or left alone, the way you would brief a human moderator. Hard filters catch known spam domains and keywords before the AI runs. Test your rules against recent comments and see what would happen before you publish.
Rules you can trust before they go live
- No regex, no decision trees
- One rule per line, in your own words: "Hide comments that mention competitor brands", "Flag anything about refunds for review".
- Preview on real comments
- Run a draft against your recent comments and see which ones would be hidden, flagged or answered. Nothing changes on your pages until you publish.
- Versioned, with rollback
- Every published version is kept. If a change does not work out, roll back to an earlier one.
- Separate rules for DMs
- Direct messages get their own rule set, so you can treat private messages differently from public comments.
Common questions
What makes a good rule?
Be specific. "Hide comments containing competitor brand names" works better than "hide bad comments". Write one rule per line or paragraph, and say what to keep visible as well as what to hide.
Do the rules apply to every page?
Yes. Rules apply to every page in a workspace. If two brands need different policies, put them in separate workspaces.
What happens when a comment matches no rule?
It stays visible. If the AI is unsure whether a rule applies, it marks the comment Needs review instead of acting.
Try it on your own pages
Connect a page and see the inbox fill with real comments within minutes.
Start free trial