What the spam system watches
Spam detection is enabled by default. The starting rules trigger on three identical messages within 30 seconds, five messages within 10 seconds, a behaviour score of five points, or five combined user and role mentions. The behaviour score has a 30-second half-life, and the default timeout is 60 minutes.
- Exact repeated messages and near-duplicates.
- Rapid bursts and a decaying activity score.
- Multiple links and character flooding.
- A configurable number of unique user and role mentions in one message, plus @everyone or @online.
Adjust protection for one server
Open the server dashboard, choose Protection Centre, then go to Spam protection. Changes apply only to that server after you select Save all changes; every other server keeps its own settings.
- Repeated-message count and window: choose the number of identical consecutive messages and the time in which they must arrive.
- Rapid-message count and window: choose the total number of messages and the shorter burst period that triggers the rule.
- Trigger score and score memory: set the score needed to trigger and how many seconds of inactivity it takes the score to halve.
- Mention limit and timeout length: set the combined user-and-role mention limit and the timeout applied after any spam rule triggers.
- Restore defaults: return every spam adjustment to the server defaults before saving.
What enforcement does
Rapid, repeated and scored spam normally produces a timeout and warning, which can also advance the server's warning policy. The triggering message is not always deleted. The mass-mention path separately attempts to delete the message before applying its moderation response.
Dashboard and command controls
Use the dashboard to change numeric thresholds, timing and timeout length. The commands enable, disable or inspect the server's saved configuration.
~spam-detection statusShow whether protection is enabled and report the server's current thresholds.
~spam-detection enableEnable the server's configured spam controls.
~spam-detection disableDisable the spam controls.
~policy showReview the warning escalation that can follow spam warnings.
How to verify it safely
- 01
Use a test account and channel
Avoid testing in a busy public channel or with a production moderator account.
- 02
Trigger one pattern at a time
Test repeats, rate and mentions separately so the resulting event is understandable.
- 03
Check the real member state
Confirm the timeout and warning actually exist instead of relying on one success message.
- 04
Review the case and event history
Use the dashboard to make sure the record matches the action that occurred.