In the evolving landscape of medical imaging, this study introduces a deep learning-based approach for brain tumour detection and classification. In this study, a U-Net architecture was developed for tumour detection and segmentation while an EfficientNet-based model was used for classification. Dataset consisting of MRI scans which has complex brain tumour pattern types was used to train the model. The performance of the developed model was evaluated using Dice coefficient, IoU score, sensitivity, and specificity for detection, and accuracy, precision, recall, and F1-score for classification, which demonstrates the system's effectiveness. The detection model achieves a Dice coefficient of 0.9321 and an IoU score of 0.8729, while the classification model attains an overall accuracy of 0.965, which surpasses the benchmark methods. Additionally, a user-friendly web interface was developed to enhance the system's practicality for clinical use. The results obtained show that the developed interface enables real-time tumour analysis. The proposed system not only improves the accuracy and efficiency of brain tumour analysis but also provides a seamless tool for medical professionals, which will enhance diagnostic workflows and patient outcomes.