Prolonged sitting in workplaces contributes to health issues like musculoskeletal disorders, making proper posture maintenance essential yet challenging. This study presents a real-time sitting posture detection system using You Only Look Once version 5 (YOLOv5) to classify postures as “Good” or “Bad” and provide instant feedback for healthier habits. A key contribution is a dataset of 1,342 images collected to capture diverse user profiles and comprehensive views of sitting postures, reflecting real-world conditions, including background noise, ensuring reliable performance even in challenging environments. The model demonstrates substantial accuracy in detecting and classifying postures by achieving a mAP@0.5 of 78.9%, a precision of 74.4%, and a recall of 79.7%. Integrating webcams and leveraging cloud platforms offers efficient, accurate, and non-invasive feedback, promoting workplace ergonomics and healthier sitting practices.