Volume 1 - Issue 4, November - December 2025

In order to address communication barriers between hearing and deaf people, this paper describes the creation of an AIbased application that enables real-time translation of sign language into speech and text. Convolutional Neural Networks (CNNs), Media Pipe for real time gesture recognition, an easy-to-use user interface, and a feedback loop for ongoing enhancement are all features of the system. The suggested system improves accessibility and inclusivity in fields like public services, healthcare, and education. Future developments will include support for more sign languages and integration with wearable technology. This paper presents an advanced AI-driven system for Real-time sign language Translation, aiming to eliminate communication barriers between deaf and hard-of-hearing (DHH) individuals and the hearing community. The proposed framework integrates Computer Vision, Deep Learning, and Natural Language Processing (NLP) to enable seamless gesture-to-text and speech-to-sign translation. Utilizing MediaPipe for real-time keypoint detection and a hybrid CNNLSTM-Transformer architecture, the system effectively captures spatial and temporal patterns of hand gestures and facial expressions. This research contributes to the development of intelligent assistive technologies that promote social inclusion, accessibility, and equality, offering a scalable path toward barrier-free human communication through artificial intelligence.