CUB Brain
Smart moderation that isn't AI, and always says why
Scores every message with our own libraries, with no AI: scams and phishing (including look-alike links), harassment and threats, telling someone to hurt themselves, sexual content, slurs (stored only as fingerprints), spam, and raids (the same message from several members at once, links from brand-new accounts). Phrases are matched in English, Spanish, Portuguese, German and French, through l33t, look-alike letters and spacing tricks. Pick a preset (Family friendly, Gaming, Strict, Streamer chat or Support server) or set your own score for each action: log, delete, warn or a 10-minute timeout. Every decision says why.
- Free, like every feature
- Off until you switch it on
- Works in the background once on
Set it up
- Switch on CUB Brain
Dashboard → your server → Modules → Message Filters → CUB Brain. It starts off, like every feature.
- Open CUB Brain in the sidebar
Its settings save as you change them and work straight away.
- Preset
Choose what suits your server.
- Log
Choose what suits your server.
- Delete
Choose what suits your server.
- Warn
Choose what suits your server.
Settings
Everything on the CUB Brain page of the dashboard. Changes save as you make them.
- Preset
- Choose what suits your server.
- Log
- Choose what suits your server.
- Delete
- Choose what suits your server.
- Warn
- Choose what suits your server.
- Time out (10 min)
- Severity: a mild match scores 1, a serious one 2, a severe one 3 (scams, threats, slurs). An account under a week old adds 1, joining in the last hour adds 1.
- Log to
- Choose what suits your server.
- Tell the member why
- The short note after a removal names the first reason