# Anti Spam

Anti Spam protects your server from spam sent by users by analyzing messages and finding patterns in them.

Once a pattern is found, all messages matching it are blocked. Patterns will expire after they haven't been triggered for a few minutes.

The following Anti Spam rules are used to combat spam, however they can cause false-positives, so you can disable them entirely or on a per-channel basis on the Dashboard.

# traffic.similar

This is the most basic defense against spam, by detecting similar messages. To trigger this rule, a has to send 5 matching messages in 30 seconds or a total of 12 matching messages per 30 seconds (ignoring who sent the message).

# traffic.exact

This rule defends against sending the same message in multiple channels. This was implemented mainly as a fallback against bots that send scam messages in every channel and aren't detected by other rules.

# traffic.token

This rule detects recurring words or phrases ("tokens") in messages. This will catch more suphisticated spam and is generally harder to trigger accidently.

# traffic.sticker

This rule will catch sticker spam: sending the same sticker 3 times in 30 seconds. This rule was created before the ratelimit on "Wave to say hi" buttons of welcome messages was introduced, and lost importance since.

# traffic.attachment

This rule will catch attachment spam: sending the same attachment 3 times in 30 seconds. For performance reasons only the filesize is compared and not the literal content.