Food spoilage and waste are critical global issues, contributing to significant economic losses, food insecurity, and environmental degradation. Traditional methods of spoilage detection, such as visual inspection or reliance on expiration dates, are inefficient, inaccurate, and lack real-time applicability. Existing systems like ImpactVision, Intello Labs, and TOMRA Food Sorting Machines have made advancements in optical food sorting, but these systems face major challenges, including high costs, technological complexity, and limited scope in detecting spoilage beyond superficial visual cues. In response to these challenges, we propose Rackit, an AI-powered solution for comprehensive and real-time spoilage detection. Rackit employs Convolutional Neural Networks (CNN) integrated with TensorFlow to detect spoilage indicators such as discoloration, texture changes, and mold growth at early stages."-->. Rackit’s Streamlit-based web application seamlessly integrates into existing supply chain processes, offering a user-friendly interface without disrupting current workflows. Quantitative analysis from simulations shows that Rackit can detect spoilage with over 90% accuracy while significantly reducing food waste. The system processes inputs in real-time, adapting to changes in environmental factors like temperature and humidity, and offers an end-to-end solution for both large-scale food producers and smaller enterprises. With future enhancements, including IoT integration and predictive analytics, Rackit has the potential to revolutionize food safety, reduce foodborne illness, and contribute to sustainable food supply chains globally.