This research presents a realtime yoga pose correction system that integrates AI driven pose estimation, similarity checking, and audio feedback mechanisms. The system’s goal is to assist users in achieving correct yoga postures by providing immediate, personalized feedback. By using machine learning and sophisticated computer vision methods, the system detects and evaluates user poses, identifying misalignments, and offering corrective guidance through audio prompts. By offering a responsive, user friendly, and interactive experience, the proposed system aims to promote safe and effective yoga practice, reduce the risk of injury, and support the development of correct form and technique over time.