How To Set Up Built-In Chat Moderation

SheevChat can detect unwanted chat messages and remove, warn, time out, or ban viewers across supported platforms.

Start Safely

  1. Open Bot Commands → Moderation.
  2. Configure one rule at a time while the main Moderation toggle is off.
  3. Use forgiving limits and appropriate Bypass role settings.
  4. Enable the individual rule, then enable Moderation.
  5. Test with a trusted account before relying on it live.

SheevChat and the relevant platform connection must remain online. The connected bot or moderator account also needs permission to delete messages, time out viewers, or ban them.

Available Rules

  • Link Guard: allow trusted domains and block unwanted domains or patterns.
  • Phrase Rules: create groups of phrases and optionally check messages or usernames.
  • Caps Guard: limits capital letters using message length, count, and percentage.
  • Symbol Guard: limits excessive symbols.
  • Emote Guard: limits the number of emotes.
  • Message Length: catches messages above a maximum character count.
  • Repeat Guard: catches repeated words or phrases.

Each rule can use Remove message, Warn + remove, Timeout, or Ban. Timeout rules also use the configured Duration.

  1. Add one allowed domain or pattern per line, such as youtube.com.
  2. Add known unwanted domains or patterns to the blocked list.
  3. Choose the action, duration, and bypass role.
  4. Optionally enable Chat Note, write a public message, and set its cooldown.
  5. Enable Repeat Offenders if repeated violations should multiply timeout duration.

The repeat-offender reset controls how long SheevChat remembers recent Link Guard violations.

Configure Phrase Rules

  1. Enable Phrase Rules and click New Group.
  2. Name the group and enter one phrase per line.
  3. Choose whether it checks chat, usernames, or both.
  4. Select its action, duration, bypass role, optional moderation reason, and chat note.

Leave Regex off unless you understand regular expressions. Enable Match case only when capitalization should matter.

Avoid False Positives

  • Begin with Remove or Warn + remove before using bans.
  • Exempt trusted roles where appropriate.
  • Keep thresholds high during initial testing, then lower them gradually.
  • Add legitimate media domains to Link Guard's allowed list.
  • Review separate phrase groups instead of placing unrelated rules together.
  • If moderation does nothing, confirm the master toggle, rule toggle, platform connection, and account permissions.