Skip to content

BruzelBrunft/SaferChat-AI

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 
 
 
 
 

Repository files navigation

Invite Bot Website Support Server Free Beta

SaferChatAI Logo

SaferChatAI

AI-powered Discord moderation that understands context — not just keywords.

Harassment, hate speech, and toxic behaviour don't follow rules. SaferChatAI catches what AutoMod and MEE6 miss — without ever auto-banning. Your moderators always decide.

→ Add to Discord — Free · Website · Support Server


Why SaferChatAI?

Keyword filters are blind to context. A message like "not being racist but statistically those communities are just more aggressive" contains no slurs, breaks no keyword rules — and passes straight through AutoMod.

SaferChatAI reads the conversation, not just the words. It flags what actually matters and sends a full report to your mod team. No auto-bans. No false bans. Just the right information, at the right time.


Features

🧠 Context-aware AI moderation — understands subtext, coded language, and escalation patterns that keyword filters miss

🛡️ Human-in-the-loop — every flagged message gets Delete, Warn, Timeout, or False Positive buttons. The AI spots it, your mods decide

💙 Crisis detection — self-harm language triggers a private, compassionate message to the user and an immediate urgent alert to your mod team

Instant word filter — custom blocklist that runs before everything else. Zero delay, zero API cost

🔗 Advanced link filtering — 5 configurable modes to control what links are allowed in your server

📊 Moderation analytics — 30-day incident breakdown, top offenders, false positive rate — all in one dashboard

🎯 Gets smarter over time — every false positive your mods mark teaches the bot. It calibrates to your server's culture automatically


How It Works

SaferChatAI uses a two-layer approach to balance accuracy and cost:

  1. Instant filters — word filter and link filter run locally with zero latency
  2. AI analysis — suspicious messages are scanned for harassment, hate speech, self-harm, sexual content, and violence
  3. Context review — borderline cases get a deeper context analysis to eliminate false positives before your mods ever see them

Result: High catch rate, low false positive rate, and reports that include the full conversation thread so your mods have everything they need.


Commands

Command Who What
/config Admin Main setup: mod channel, auto-delete, link mode, exempt roles/channels, word filter
/advanced Admin Per-category thresholds (Strict / Standard / Relaxed / Off), self-harm ping settings
/mod history @user Moderator Full incident and action history for a user
/mod warn @user Moderator Send an official warning via DM
/statistics Moderator 30-day overview: incidents, false positive rate, top offenders
/help Everyone Commands and setup guide

Setup — 2 Minutes

1. Add SaferChatAI to your server

2. Run /config in any channel — a visual panel opens. Select your mod channel and you're done.

3. SaferChatAI starts monitoring immediately. Reports appear in your mod channel with one-click actions.

No configuration files. No rules to write. No credit card.


Permissions

SaferChatAI requests only what it needs:

Permission Why
Read Messages To analyse messages
Send Messages To post reports to your mod channel
Manage Messages To delete content when your mod approves it
Moderate Members For the Timeout button

SaferChatAI never auto-bans and never reads DMs.


Privacy

Messages are sent to a third-party AI provider for analysis. They delete API data within 30 days and do not train on API inputs. Only flagged messages are stored in SaferChatAI's database. Normal conversation is never saved.

Full Privacy Policy


Free Beta

SaferChatAI is fully free during the beta period. No credit card, no trial expiry, no feature gating.

When beta ends: a free tier remains free forever. Beta users get early access and a permanent discount on paid plans.

Add to Discord — Free


About

AI-powered Discord moderation bot that understands context — catches harassment, hate speech & toxic behaviour that AutoMod misses.

Topics

Resources

Stars

Watchers

Forks

Contributors