This paper introduces a comprehensive grading system to enhance product quality control in industrial and manufacturing processes. The system includes live color detection, damage identification, quality classification, and robotic arm manipulation. Live color detection allows real-time analysis of product colors, while damage detection uses computer vision algorithms to identify defects. Quality classification uses machine learning techniques to categorize products based on predefined parameters. The SVM model, with RBF kernel, was found to be the most effective for quality classification, achieving an accuracy of 94.78% for oranges and 84.77% for pallets. YOLOv8 was used for defect detection, with the most successful run resulting in 88.55% precision, recall, F-Score, and mAP at 50.70%. The system also integrates a robotic arm for real-time manipulation and removal of detected defects. FlexiGrade, the general-purpose automatic product grading system, is a versatile framework that can be adapted and enhanced for various products, making it a robust and scalable solution for different needs.