OrphanGuard AI

AI-powered transparency and accountability platform that monitors orphanage environments for violence, emotional distress, and safety violations to protect vulnerable children in real-time.

Comment

Inspiration

On September 20, 2026, I visited an orphanage in Lahore, Pakistan with my university's GMS Society. What I witnessed was deeply disturbing — workers beating children, no accountability systems in place, and 8-year-olds displaying unnaturally mature behavior from years of unchecked abuse. Pakistan has 4.6 million orphans, and most orphanages operate with zero transparency or oversight.

I decided to build a technology solution that could bring accountability to these environments and protect the most vulnerable children.

What it does

OrphanGuard AI is a real-time monitoring and accountability platform for orphanages that uses AI-powered detection to identify:

  • Violence & harassment — 8 detection categories including physical violence, verbal abuse, inappropriate contact, and neglect
  • Emotional distress — Identifies children showing signs of distress, fear, or crying through behavioral analysis
  • Security anomalies — Headcount mismatches, perimeter breaches, and missing child alerts
  • Health & growth tracking — Children's registry with health records, height/weight/BMI monitoring over time

The platform provides a multi-orphanage dashboard where authorities can monitor multiple facilities simultaneously, with orphanage ranking scores (0-100, A-F grades), anomaly detection for week-over-week trends, visitor management with CNIC verification and webcam photo capture, and daily briefing reports.

How we built it

  • Backend: Node.js + Express with WebSocket for real-time updates
  • Frontend: Vanilla JavaScript with responsive dashboard UI
  • AI Detection: Computer vision pipeline for violence detection (ViolenceNet v2.1), emotion recognition, and behavioral analysis
  • Data: 10 seeded Pakistani orphanages with real institutional names and cities
  • Deployment: Railway (auto-deploy from GitHub)
  • Notification System: Real-time WebSocket push notifications for critical alerts

Challenges we ran into

  • Designing detection categories that cover the spectrum of child abuse without false positives
  • Building a real-time system that can process multiple camera feeds simultaneously
  • Balancing surveillance capabilities with children's privacy and dignity
  • Creating a scoring algorithm that meaningfully ranks orphanage safety

Accomplishments that we're proud of

  • Built a fully functional multi-orphanage monitoring dashboard
  • Designed 8 distinct violence/harassment detection categories based on real observed abuse patterns
  • Created an orphanage ranking system that incentivizes better care
  • Deployed a working prototype accessible to potential NGO and government partners

What we learned

  • The scale of the orphanage accountability gap in developing countries
  • How AI can be applied thoughtfully to child protection without becoming invasive surveillance
  • The importance of designing for the specific cultural and institutional context (Pakistani orphanage system)

What's next for OrphanGuard AI

  • Pilot deployment with Punjab Social Welfare Department and Child Protection Bureau
  • Integration with RFID systems for automated check-in/check-out
  • Mobile app for field inspectors and social workers
  • Partnerships with Edhi Foundation, Al-Khidmat, and SOS Children's Villages
  • Grant applications to UNICEF Venture Fund and EU child protection programs
  • Expanding to cover all 4.6 million orphans across Pakistan's orphanage network

Built With

Share this project:

Updates

Submission history