Moderation

Configure auto moderation

Detect repeated messages, mentions, emojis, blocked words and invites with tailored rules.

On this page
Dashboard → Automodération → Configure auto moderation

In pictures

Auto moderation rules
Auto moderation rulesOriginal image

Before you begin

The rules.manage permission and permissions for the selected action are required. Text analysis depends on message data available to Spider.

Step by step

  1. Open Rules and sanctions → Auto moderation and create a rule named after its purpose.

  2. Choose a detection type and configure its threshold, words or expression.

  3. Set the affected groups, monitored channels and role or member exclusions.

  4. Choose the response, a sanction template or the available heat behaviour. Test exceptions before enabling the rule.

  5. Save, enable the module and review recent triggers.

Choose the right detection

  • Repetition: adjust counts and the observation window to detect repeated identical messages.
  • Mentions and emojis: limit abuse without disrupting normal conversation.
  • Words and expressions: start with simple matching and carefully review regular expressions.
  • Invites: define where and which roles are allowed to share invitations.

Remove the entire burst

When deletion is enabled for a repetition rule, Spider removes the matching messages that reached the threshold, including those sent before detection. Deletions are grouped per channel and messages already handled by another protection are not deleted again. A punishment cooldown prevents penalising every message in the same burst while cleanup continues.

Avoid duplicate responses

Check overlap with Messages and scams and link protections. A rule applying an explicit template does not need a second manual punishment for the same event. Exclusions and target groups are configured per rule.

Practical example

For a public channel: detect repetition within a short window, remove affected messages and begin with a warning. Exclude a command channel where similar messages are expected.

Questions and solutions

The rule triggers too often

Review matching mode, threshold and affected channels. Test words in normal contexts and add targeted exceptions rather than disabling all protection.

All documentation