Inspiration

FatigueGuard was inspired by the long hours students and developers spend in front of screens, especially during exams, hackathons, and work sessions. We noticed that people often ignore their posture, eye strain, and overall well-being while trying to stay productive. This can lead to fatigue, burnout, and long-term health issues. We wanted to create a solution that actively helps users stay healthy while working, without requiring extra effort or awareness from them.

What it does

FatigueGuard is a real-time wellness assistant that uses computer vision to monitor posture, eye strain, and user activity. It provides live feedback and alerts to encourage better habits, such as sitting upright, blinking regularly, and taking breaks. The system also supports SMS alerts for important reminders, making it useful even when users are deeply focused.

How we built it

We built FatigueGuard as a web-based application using browser-based computer vision to analyze user behavior in real time. We integrated Supabase for authentication and backend services, and Twilio for SMS notifications. The application processes webcam input directly in the browser and uses lightweight logic to determine posture and fatigue-related patterns. The UI was designed as a clean dashboard to provide clear and actionable feedback.

Challenges we faced

One of the biggest challenges was configuring third-party services like Google OAuth and ensuring all redirect URIs were correctly set up. Even small mismatches caused authentication failures. Integrating Twilio also required understanding trial limitations and properly formatting phone numbers. Additionally, working with real-time webcam data required balancing performance and responsiveness while keeping the experience smooth for users.

What we learned

We learned how to integrate multiple APIs and services into a cohesive product under time constraints. We also gained experience debugging authentication flows, handling real-time data in the browser, and designing user-focused interfaces. Most importantly, we learned how to prioritize features and focus on delivering a polished, working demo.

Built With

Share this project:

Updates