This paper presents a real-time, vision-based system for Indian Sign Language (ISL) recognition and translation, aimed at enhancing communication between the deaf community and non-signers. The system combines a CNN-LSTM architecture for static gesture recognition, achieving an accuracy of 98.47% and introduces GestureNet, a bidirectional LSTM model trained on a custom dynamic gesture dataset, which attains 96.83% recognition accuracy. Additionally, a Generative AI framework is integrated to convert recognized gestures into semantically coherent and contextually appropriate sentences. By emphasizing real-world applicability and high recognition performance, the proposed system advances sustainable and accessible communication technologies, with potential impact in education, public services, and digital inclusion, particularly in developing regions.