A real-time system for translating sign language into text has been developed to bridge communication barriers for the Deaf and Hard of Hearing (DHH) community. The model utilizes the Inception V5 deep learning architecture, enhanced with additional dense, batch normalization, and dropout layers to achieve an accuracy exceeding 90%. The system captures gestures via a camera feed, preprocesses them, and recognizes the gestures with high accuracy, displaying the output as readable text to provide smooth communication. Data augmentation techniques, including rotation and zoom, enhance the model’s robustness in various lighting and background conditions which further ensures reliability in real-world applications. With real-time processing capabilities, the system provides instant feedback, making it adaptable for everyday interactions. Future advancements could involve extending support for multiple languages and additional gesture classes, further enhancing accessibility and inclusivity in communication.