Abstract: Medical image classification is an important aspect when it comes to early detection and diagnosis of various diseases, helping medical practitioners make informed decisions. This research proposes the application of different deep learning algorithms for the classification of medical images related to brain tumors, tuberculosis (TB), and fractured bones, using CNNs. A web interface based on Flask has been created to enable automatic diagnosis by allowing users to upload medical images. The CNN models have been trained with labeled datasets and optimized to achieve high accuracy in real-world cases. The approach for the research contains details about the dataset preprocessing, model architecture, training philosophies, and metrics for performance evaluation, such as accuracy, precision, recall, and F1-score. The findings show a vast enhancement in classification accuracy when juxtaposed with existing manual mechanisms and pave the way for keeping deep learning for medical diagnosis.