🛡️ AI Moderator
The AI Moderator tab is one rules page for everything the AI watches for in a Telegram group: its own policy judge, the built-in detectors, alert rules, your custom rules, and new-member checks. This page covers Telegram only; Discord's AI Features are a different page.
Where to find it
- Open your workspace in the Fren One webapp.
- In the sidebar under Channels, select the Telegram group. Its pages open beneath it.
- Select AI Features, then the AI Moderator tab.
Switches and levels save instantly on this page. Opening a rule shows its full settings, which save from that dialog. There's no page-wide Save bar.
The moderator itself
Let the AI Moderator read messages and enforce your community rules is the master switch. With it on, an AI judge reviews every message against your community's moderation policy and acts on violations. Every action is posted with a one-tap Revoke, and each revocation teaches the judge.
- Default autonomy sets how the moderator handles violations in categories without their own level: Act automatically enforces above the confidence bar, Suggest to admins proposes the action with one-tap approval, and Observe only records verdicts with no actions or alerts.
- By default it acts on its own only above a per-action confidence bar (warnings 80%, deletions 85%, mutes 90%, kicks 95%, temporary bans 97%) and turns anything less certain into a suggestion. It never bans anyone permanently: a ban verdict becomes a temporary ban plus a confirm button for your admins.
- Settings opens the moderator's settings: where notices and suggestions are sent, mute and temporary-ban durations, and whether messages are judged one by one and whether conversation windows are reviewed for patterns. They save with Save settings.
- Policy shows the active policy pack and version. Change policy lets you apply a ready-made pack (Generic, Crypto community, or Family-friendly) or write your own rules in plain language for the AI to compile; nothing is saved until you confirm what it understood.
When the master switch is off, policy and alert rules are paused. Built-in filters and your custom rules keep running on their own switches.
The rules
Rules are grouped by what they look at. Each shows where it comes from:
| Group | Rule | Type | More detail |
|---|---|---|---|
| Messages | Fast spam filter | Built-in | Spam Detection |
| Messages | Hostility | Built-in | Sentiment Moderation |
| Messages | Spam, Scams & phishing, Harassment, NSFW, and any others your policy adds | Policy | Set by the policy pack or your own rules |
| Messages | FUD, Urgent issues | Alert | Judged over the recent conversation; need chat monitoring on the Alerts & Reports tab |
| Your custom rules | Your own AI rules, managed with Manage custom rules | Custom | Smart Triggers |
| New members | Suspicious profile | Built-in | Member Screening |
| New members | Impersonation | Built-in | Impersonation Detection |
Each row has a switch or level control that saves instantly. A policy rule can Inherit the default autonomy or set its own level, and a rule whose suggestions were all approved may offer Ready to act to switch it to acting automatically. Opening a rule shows everything else: what it detects, its response, sensitivity, and for built-in rules the full feature settings.
Alerts & Reports
The Alerts & Reports tab is where the moderator's alerts are delivered, separately from the rules:
- Monitor the chat for issues that need admin attention, with What to watch for → Unanswered questions, and where monitoring alerts go. FUD and urgent-issue alerts moved to the AI Moderator rules above.
- Where do screening alerts go?: the destination for new-member screening alerts.
- Send periodic community reports: a report on a schedule, and where it's sent.
Each section there saves with the Save bar.
Related features
- ✨ AI Assistant: Answers member questions, a separate tab
- 🧠 Smart Triggers: The custom rules
- 🔎 Member Screening: The Suspicious profile rule