AI Discord
Moderation.
KenwoodBot uses AI to understand message context — not just keyword lists. 92% catch rate, ~3% false positives. Moderation that actually works.
Why AI beats
keyword filters.
Traditional keyword filters match against word lists. They catch ~65% of violations with ~15% false positives. AI moderation reads message context, detects sarcasm, obfuscated language, and multi-message harassment patterns. KenwoodBot's AI catches ~92% of violations with ~3% false positives — a 40% improvement.
| Metric | Keyword Filter | KenwoodBot AI |
|---|---|---|
| Toxicity catch rate | ~65% | ~92% |
| False positive rate | ~15% | ~3% |
| Sarcasm handling | Poor | Strong |
| Obfuscation handling | Poor | Strong |
| Multi-message context | None | Yes |
| Setup time | Hours | Minutes |
| Latency per message | <50ms | 100-250ms |
Everything you need
to keep your server clean.
AI Content Moderation
Context-aware AI detects toxicity, harassment, spam, and threats using LLMs via OpenRouter. Understands sarcasm and obfuscated language.
Auto-Mod Rules
Keyword filters, regex patterns, spam detection, mention limits, and raid protection. Multi-layer defense stack.
Sanction System
Warn, mute, timeout, kick, ban — with duration, reason, and automatic expiration. Timed punishments that clean up after themselves.
Casier (Mod History)
Complete moderation history per member. View all sanctions, notes, and incidents in one place.
Lockdown
Instant channel or server lockdown in case of raid or crisis. One command locks everything down.
ModQueue
AI-assisted moderation queue. Flagged messages queued for human review. Your team reviews, KenwoodBot handles the volume.
Moderation commands.
Every moderation action available as a slash command. Context menu support for quick access.
/sanctionApply warn, mute, kick, or ban with reason and duration
/casierView full moderation history for any member
/modnoteAdd private notes visible only to staff
/lockdownLock or unlock channels instantly
/banBan a member with reason and evidence
/kickKick a member from the server
/muteTimeout a member temporarily
/unbanReverse a previous ban
/warnIssue a warning to a member
/modqueueView and manage the moderation queue
/intelFull intelligence suite for your server
/radarCommunity radar and activity analysis
Frequently asked
questions.
How does AI moderation on Discord work?
AI moderation runs every incoming message through a language model trained to detect toxicity, harassment, spam patterns, and harmful intent. Instead of matching word lists, it evaluates the meaning of a message — including sarcasm, obfuscated language, and multi-message context.
Is AI moderation better than keyword filters?
Yes. AI moderation reaches ~92% accuracy with ~3% false positives, compared to ~65% accuracy and ~15% false positives for keyword filters. AI also catches sarcasm, veiled threats, and creative misspellings that keyword lists miss entirely.
What moderation actions does KenwoodBot support?
KenwoodBot supports warn, mute, timeout, kick, ban, lockdown, and moderation notes. Each action can include a reason, duration (for temporary actions), and automatic logging to the member's casier (moderation history).
Does AI moderation work in real-time?
Yes. KenwoodBot processes messages in 100-250ms, which feels instantaneous to users. Flagged messages are removed before most users see them in active channels.
Can I combine AI moderation with keyword filters?
Yes, and we recommend it. The best moderation stack is AI + keyword filters + rate limits + human review. This layered approach delivers ~95% catch rate with ~2% false positives.