AutoMod
Automatic content filtering with strikes, escalation, anti-raid and NSFW image detection.
AutoMod watches messages in real time and acts on rule violations on its own - deleting, warning, timing out, kicking or banning - while tracking strikes that escalate repeat offenders.
Rules
Eight rule types can be enabled and tuned independently:
- Banned words - block a configurable word list
- Spam - rapid repeated or duplicated messages
- Links - block links (with an allowlist)
- Discord invites - block invite links to other servers
- Caps - excessive capital letters
- Emoji - too many emojis in one message
- Attachments - block disallowed file types
- NSFW images - flag explicit images via OpenAI moderation
Each rule has its own action (delete / warn / timeout / kick / ban), optional duration, and strike points.
Strikes & escalation
Every violation can add strike points to the offender. Points decay after a configurable number of days. When a user’s total crosses a threshold you define, an escalation action fires automatically - for example 5 points → 1 h timeout, 10 points → ban.
Anti-raid
A burst of new joins triggers a raid alert and optional lockdown, which auto-kicks new members until you clear it.
NSFW image detection
The NSFW rule sends image URLs to OpenAI’s moderation endpoint. Add your API key (stored encrypted), pick the categories to flag, and set a confidence threshold.
Lower threshold = stricter
Exemptions & logging
Exempt trusted roles or channels globally or per rule. Set a log channel to receive a record of every action and strike.