AI-Powered Crisis Detection
Protecting Lives Through
Real-time NLP analysis detecting mental health crisis signals with clinical-grade severity scoring — all running privately in your browser.
🧠
Multi-Layer NLP
6-category crisis lexicon with sentiment analysis, urgency detection, and protective factor scoring
🔒
100% Private
All analysis runs in-browser. Zero data ever sent to external servers. Privacy by design.
⚡
Real-Time
Instant analysis with live monitoring streams and automated severity-based alert routing
📋
Response Protocol
Automated clinical-grade response recommendations based on severity classification
Crisis Resources
If you or someone you know is in crisis, help is available 24/7
⚠️ CrisisGuard AI is a detection tool, NOT a replacement for professional mental health services.
About CrisisGuard AI
Built for UOE Summer of Code 2026
🎯 The Problem
Millions of crisis signals go undetected in digital communications. Organizations lack tools to proactively identify at-risk individuals.
💡 Our Solution
Multi-layer NLP pipeline analyzing text for crisis indicators using sentiment analysis, contextual patterns, and severity scoring.
🏗️ Architecture
Client-side inference, multi-stage pipeline, real-time analytics, PWA architecture. Zero external dependencies for core analysis.
🔒 Privacy First
All analysis in-browser. No data sent externally. No storage. People in crisis deserve protection, not surveillance.
Tech Stack
HTML5CSS3Vanilla JSChart.jsNLP EngineSentiment AnalysisCanvas ParticlesClient-Side ML