Moderation overview

Moderation is opt-in per Watcher — nothing is enabled without your choice, and it's independent of media safety, ordinary conversation, and Auto-Posting.

A Watcher can act on: unsolicited promotion and repeated advertisements, phishing and scam links, link spam/flooding, and explicit content — both in text links and in photos/videos where media moderation is enabled. See "Natural Moderation" and "Media Moderation" for how each is handled.

For unsolicited promotion and link spam, enforcement is graduated: a warning first, escalating only on repeat offenses within the same Community — never an immediate ban for a first-time, lower-severity violation.

For clearly severe cases (explicit content, phishing/scam links at high confidence), enforcement is immediate — these are treated as safety issues, not subject to the graduated-warning path.

Available actions depend on what the Watcher's Telegram bot account was actually granted: deleting the message, warning the user, and — where the bot has admin rights to restrict or remove members — escalating enforcement on repeat offenses. See "Permissions" for exactly what's required for each action.

Every real moderation decision — warn, mute, kick, or purge — is recorded in that Community's Activity feed, and (where a private Escalation Group is connected) reported to your team in a structured Watcher Ops report — see "Watcher Ops Moderation Reports." Routine chatter and silence are never logged as moderation events.