Blog
Notes from building Civora.
Long-form on what works in modern Discord moderation — context-aware AI, multi-language, EU compliance, and the gap between what marketing claims and what the bot actually catches.
- 5 min read
Why keyword bots miss 60% of toxic messages (and what to do about it)
MEE6, Carl-bot, and AutoMod all share the same blind spot: they look at words, not meaning. Here's the math on why that matters — and what context-aware moderation actually catches.
ModerationAIRead article - 7 min read
GDPR-compliant Discord bots in 2026: a moderator's checklist
If you run a Discord server with EU members, the bot reading their messages is your sub-processor. Here's the practical checklist for staying clean.
GDPRComplianceRead article - 6 min read
Hungarian Discord moderation: why most bots fail at it (and what works)
Hungarian morphology, suffix chains, and the sheer rarity of HU in training data make it a stress test for any moderation model. Here's what actually works.
HungarianModerationRead article