Sign language is crucial for those with hearing and speech impairments, enabling them to communicate effectively with others and express themselves. Therefore, it is necessary to create a sign language recognition (SLR) system for the interpretation of sign language. Numerous sign languages are used globally, with American Sign Language (ASL) being the most widely used. This paper reports the design of a real-time ASL recognition system using methods from machine learning and computer vision. The system uses MediaPipe’s hand feature extraction library and a convolutional neural network (CNN) classifier. The Teachable Machine web tool is utilized to train the dataset for the model. The system translates the hand shapes of the ASL fingerspelling alphabet into corresponding English text letters. The proposed system has achieved 99% accuracy for 26 letters, and this result indicates that it could be beneficial in enhancing the quality of life for individuals who are deaf or speech-impaired.