The recognition and interpretation of American Sign Language (ASL) using computational methods stands as a significant advancement in bridging communication gaps between the Deaf community and the hearing world. This project aims to develop an accurate and By employing deep learning techniques, we propose a model that interprets ASL signs from static images and video sequences, enabling seamless translation of sign language gestures into textual or spoken language.Our methodology encompasses the collection and preprocessing of a comprehensive dataset consisting of ASL signs, including both alphabet and commonly used phrases. Utilizing state-ofthe-art CNN architectures, the system undergoes rigorous training and validation phases to ensure high levels of accuracy and efficiency. Robustness and dependability are assessed by gauging the model's capacity to correctly identify a broad spectrum of ASL signals in a variety of backgrounds and environments.