Assess.ai

Inspiration 💡

The pandemic highlighted the difficulties educators faced in preserving academic integrity during remote assessments. Existing proctoring tools were often either overly intrusive, requiring heavy software installs, or ineffective at preventing dishonest practices. We set out to build a privacy-first, installation-free solution that keeps assessments fair, scalable, and accessible to all.

What it does 🎯

Assess.ai is a browser-based proctoring platform that:

Monitors candidate behavior in real time through in-browser facial and activity tracking
Detects multiple faces, tab switching, and other suspicious behavior
Instantly alerts instructors to potential violations
Runs entirely in the browser — no downloads or installations required
Processes all monitoring data locally to protect privacy
Seamlessly scales to thousands of concurrent test-takers

How we built it 🛠️

Frontend

React.js + TypeScript for a responsive, reliable UI
face-api.js for browser-based face detection
WebRTC for live video handling
Tailwind CSS for a clean, adaptive design

Backend

Node.js + Express for secure API endpoints
MongoDB Atlas for event logging and secure storage
Socket.io for real-time alerts
JWT for authentication

Challenges we ran into 🤔

Browser performance: Ensuring smooth, low-latency monitoring without overloading devices
Real-time processing: Minimizing detection delays and handling unstable network conditions
Privacy compliance: Keeping sensitive data entirely on the client side while still ensuring security