All notable changes to SaferChatAI will be documented here.
- Context-aware AI moderation — detects harassment, hate speech, and toxic behaviour that keyword filters miss
- Two-stage analysis pipeline for high accuracy and low false positive rate
- Self-harm detection with private user messaging and urgent mod alerts
- Advanced link filtering with 5 configurable modes
- Custom word filter with instant blocking before AI analysis
/config— visual setup panel for mod channel, sensitivity, exempt roles and channels/advanced— per-category thresholds (Strict / Standard / Relaxed / Off)/mod history— full incident and action history per user/mod warn— official warning via DM/statistics— 30-day dashboard with incident breakdown and false positive rate- One-click mod actions: Delete, Warn, Timeout, False Positive, Ban
- False positive learning — bot calibrates to server culture over time
- Multi-language website support (EN, DE, ES, FR, PT)