FlexRight

FlexRight uses AI skeletal tracking to provide realtime form feedback via webcam. Monitor reps and joint angles for a safer recovery.

FlexRight addresses the gap in home-based physical therapy: the lack of professional supervision to ensure correct form. We transformed a standard webcam into a digital coach that provides real-time feedback on exercise execution, ensuring safety and efficacy for patients recovering at home.

The platform utilizes Ultralytics (YOLOv8) to map 17 skeletal keypoints and OpenCV to calculate joint angles and render a live HUD. These libraries allow FlexRight to track range of motion and rep counts instantaneously. To ensure long-term progress tracking, all session data is synced via PyMongo to a secure MongoDB cloud database, allowing healthcare providers to review recovery trends remotely.

Built within VS Code using Gradio for the interface, the development process involved a steep learning curve in real-time data processing. Our primary challenge was integrating the AI tracking, cloud database, and UI into a single, cohesive system. This hurdle ultimately strengthened our understanding of modular design and collaborative engineering.

FlexRight has evolved from a concept into a functional prototype capable of tracking four core exercises on consumer-grade hardware. Our next steps involve expanding the skeletal math library to include more complex movements and modernizing the UI for better data visualization. FlexRight makes precision recovery accessible, data-driven, and safer for everyone.