Moderation
Scans the 200most recent draft/approved replies and proactive posts across every user against two lists: each author’s own guardrails, and the platform-wide terms below. Same substring-match logic the reply inbox uses live, just run across everyone.
Platform moderation terms
Checked against every user’s generated content, platform-wide — separate from each user’s own guardrails.
Flagged content
A slip-through here means the AI said something the author (or you) explicitly ruled out — worth a look, not necessarily wrong.
Nothing flagged in the recent scan window.
Recent admin actions
demo@personaproxyai.com update settings platform_settings — Default authenticity level set to 39/29/2026, 5:00:11 AM