The tremendous expansion in movie production and consumption has produced a major demand for automation in content moderation systems. This work intends to aid the censor board in content moderation by constructing a machine learning-based model to rate the content efficiently by minimizing time limitations. The model is meant to detect foul language and includes various advanced machine learning techniques: convolutional neural network (CNNs), Recurrent neural networks (RNNs) for processing audio characteristics, and natural language processing (NLP) approaches for evaluating subtitles. By training the model on a diverse dataset containing bloodshed and non-bloodshed images for video content and leveraging open-source data for profanity detection, such as foul comments from Twitter, we ensure a strong performance. Profanity analysis is performed using Long Short-Term Memory (LSTM) networks to identify offensive language with high accuracy. Performance indicators, including precision, recall, and F1-score, are employed to evaluate the system’s effectiveness, this novel approach provides time stamps for sensitive scenes and, thus, enables the fast search for particular frames with violent or obscene expressions. The approach significantly reduces human labor, thus increasing the efficiency of the procedure of censorship and ensuring the compliance of films with adequate viewing requirements. The deploy ability and operational readiness of the system are demonstrated by the thorough analysis and supplementary data.