Docs/Stream/Guides/Chat Policies & Moderation

Chat Moderation Policies & Runbooks

Learn how to implement automated anti-spam throttling, keyword blacklists, AI toxicity filters, and real-time moderation escalation procedures.

Automated Chat Rules

Slow-Mode Throttling

Rate limits users to 1 message every N seconds (default: 5s). Enforced across Redis clusters so reconnecting does not reset the timer.

Duplicate Suppression

Prevents spam bots from pasting repeated text by hashing each user's last 3 messages and dropping matches within a 60-second window.

Link & Caps Filter

Automatically strips or flags messages containing unwhitelisted hyperlinks or excessive uppercase characters (>70% caps).

Moderator Action Runbook

Mild Disruptions (Spam/Off-topic): Issue a timeout for 300 seconds (5 minutes) and delete the disruptive messages via delete_message.
Persistent Trolling: Issue a permanent chat ban. The user retains video watch privileges but cannot broadcast messages.
Severe Harassment / Unauthorized Leaks: Issue a stream_ban. This immediately tears down their WebRTC video feed, evicts their WebSocket connection, and records their IP/fingerprint to deny re-entry.
Chat Raid / Extreme Chaos: Execute toggle_lock with locked: true to freeze chat for non-moderators while moderators clear offending messages.
← Webhooks & EventsGeo-Fencing & IP Rules →

We use cookies and similar technologies to measure traffic and improve the site. You can choose which categories to allow. Manage Preferences.