Tomatoes are among the most important agricultural crops consumed and exported globally, necessitating rigorous quality control to meet export standards. Manual sorting methods are slow, costly, and prone to error, leading to significant inefficiencies and waste. This study introduces an affordable, automated, intelligent image-based system for tomato sorting, designed and tested to align with Jordanian export standards. Using the YOLOv8 model, a state-of-the-art real-time object detection algorithm, the system classifies tomatoes into three categories: unripe, rejected, and healthy. Classification is based on images captured from multiple angles using high-resolution cameras mounted above a conveyor belt. The system achieves a 91% accuracy rate while processing an average of 15 tomatoes per minute under varying lighting and conveyor speeds. This affordable prototype demonstrates the potential of AI-powered solutions in agriculture to enhance efficiency, reduce waste, and meet stringent quality standards. Future work will focus on increasing processing speeds and expanding dataset diversity to improve system performance.