Cyberbullying is a major concern in the digital age, especially for teenagers. The increased accessibility and usage of social media platforms have widened the issue, with incidents of cyberbullying in English experiencing a substantial rise over the past decade. This system provides a solution involves advanced AI to detect inappropriate content and involves parents in real-time monitoring. This approach embeds strong deep learning model to recognize content like hate speech, aggression, insults, toxicity, etc. It will create a safe online environment for children by detecting such contents. The architecture of this application is divided into four modules: Child, Parent, ML Model and Admin. This allows teens to share content from multiple sources that ML model will analyses for harmful content. The Parent module provides transparency and security by allowing parents to monitor their children's online activities and predictions made by the ML model. An alert mechanism notifies parents via email about cyberbullying instances. The Journal/Diary function offers a platform for teens to express their feelings. The admin module performs user management and ensures the integrity of the ML model, which is constructed with TensorFlow and Keras with an LSTM architecture, classifying cyberbullying into six classes: race, gender, ethnicity, sexuality, religion. After training on a diverse dataset, the model achieved an accuracy of approximately 94% on the training set. Project demonstrates, the potential of combining AI and web technologies to foster safer online spaces. By offering real-time alerts, engaging parents in monitoring, it embodies a holistic approach to tackling cyberbullying.