This project pioneers the automation of censor label generation in visual content by leveraging advanced computer vision and machine learning algorithms to identify sensitive scenes such as alcohol consumption and riding without a helmet in videos accurately. The system, utilizing deep neural networks trained on diverse datasets, adapts to evolving patterns and contextual variations, ensuring precise detection and alignment with ethical content standards. With its open-source framework and customizable, context-aware approach, this innovative endeavor significantly advances content moderation, contributing to a safer digital media landscape by autonomously identifying and labeling potentially objectionable scenes.