With the advancement of artificial intelligence, the quality assessment of medications or pills before they become final goods has gained more attention. This research presents a detection method for evaluating pill quality throughout the production process. The solution works with the YOLOv8 deep learning network to automate the inspection process. The dataset is trained using photos of tablets under various lighting conditions. The system's accuracy reaches 95%. The experimental findings show that the suggested model is not only accurate in defect identification, but it can also be used successfully in real-time industrial classification tasks. This approach sets the framework for developing practical applications in pharmaceutical production facilities for pill quality checking.