The early and accurate detection of brain tumors is very critical in the life of saving a patient's life and guiding proper treatment strategies. The paper discusses an automated brain tumor detection and classification system based on Convolutional Neural Networks (CNN) with architecture of VGG16. The system processes the MRI brain scans to categorize tumors in four classes: Glioma, Meningioma, Pituitary, or No Tumor. Such a study is thus supplemented by the presentation of user-friendly web applications that try to bridge the divide between medical imaging and everyday clinical use. Experiments prove to be very accurate, precise, and having great recall upholding the strength of a method that applies deep learning against traditional regulations, now with regard to overcoming its limitations. Such work improves neuro-oncology from the viewpoint of efficiency and accuracy of tumor diagnostics to global health delivery scalability.