Blify Mod is an AI moderation bot that actually reads your chat. Rather than matching messages against a list of banned words, it uses Claude (Anthropic’s AI) to evaluate them in context — so it can tell the difference between friends giving each other a hard time and someone genuinely being cruel, or between a heated debate and real harassment.
What it does
Every message in your moderated channels is evaluated against your server’s rules. When something crosses a line, Blify Mod rates its severity on a 1-to-5 scale and responds accordingly — deleting, warning, timing out, kicking, or banning as the situation calls for. Timeouts scale with severity: a rude comment might earn five minutes, while something truly egregious earns a day.
Every action is logged to a channel of your choice, and each log entry includes a button your moderators can click to reverse it. No digging through audit logs, no debating whether the bot overreacted — one click and it’s undone.
You can use the built-in ruleset (covering the usual ground — harassment, slurs, spam, NSFW content, and so on) or write your own with /setrules. If you try to add a rule that’s essentially a duplicate of an existing one, the bot will catch it and let you know.
Need a second opinion on a specific message? Reply to it and ping the bot — it’ll evaluate that message on its own. Want it to assess what’s happening in a channel right now? Ping it without a reply and it’ll review recent messages.
A few other things worth knowing
- Trusted roles or specific users can be marked as immune, so the bot leaves them alone
- A strictness setting (low / medium / high) lets you control how aggressively it intervenes
- A per-user cooldown ensures one person’s spam doesn’t quietly drain your API budget
- It plays well with Discord’s built-in AutoMod — if AutoMod already handled something, Blify Mod won’t pile on
- It’s disabled by default when you first add it, because nobody wants a bot that starts swinging before you’ve had a chance to configure it
Getting started
Setup takes about a minute:
- Add the bot, then drag its role above anyone it should be able to moderate
- /setmodchannels — select which channels it should watch (up to 3)
- /setlogchannel — choose where logs should appear
- /enable
That’s all. Run /setup-guide for a full walkthrough, or /modconfig at any time to review your current settings.
Free vs. Pro
Free is plenty for most servers — 5 custom rules, 3 watched channels, and moderation powered by Claude Haiku alongside OpenAI GPT-4o Nano. Pro raises the limits to 20 rules, adds deeper context on manual scans, and upgrades to Claude Haiku alongside Claude Sonnet — which handles the trickier judgment calls more reliably. Run /pro info for full details.
Your data
The bot only reads messages in channels you explicitly tell it to watch, or in conversations where you @mention it directly. Messages are sent to Anthropic for evaluation, but nothing is stored long-term beyond the log entries you can view yourself with /modlogs. If you ever remove the bot from your server, everything it knows about that server is automatically wiped from the database.
Why AI moderation?
Keyword filters are blunt instruments. They’ll flag “this meeting is killing me” and miss a genuine threat spelled out with invisible characters. They have no sense of tone, context, or the difference between friends and strangers. A human moderator would catch all of that — and Blify Mod aims to moderate more the way a human would, just without needing sleep.
Add it to your server and give your mod team a well-earned break.
