This paper presents the design and development of an Artificial Intelligence (AI)-based, real-time Speech-to- Indian Sign Language (ISL) Translation System that bridges the communication gap between the hearing population and the deaf and hard-of-hearing community. The proposed system captures spoken audio via a standard microphone, converts it to text using cloud-based Automatic Speech Recognition (ASR), applies Natural Language Processing (NLP) techniques including tokenization, stop-word removal, and sentence sim- plification, and maps the resulting tokens to pre-recorded ISL GIF animations for sequential visual display. Implemented as a software-only Flask web application, the system requires no specialized hardware such as sensor gloves or motion-capture devices, enabling deployment on any standard browser-equipped device. A modular pipeline architecture ensures that individual components can be independently upgraded without disrupting the overall system. Experimental evaluation demonstrates high gesture-mapping accuracy for common vocabulary in controlled conditions, with near-real-time response latency suitable for conversational use. Out-of-vocabulary words are handled via a finger-spelling fallback mechanism, ensuring uninterrupted communication. Results confirm that an accessible, cost-effective, and scalable speech-to-sign translation tool can be realized using widely available web technologies and AI-driven NLP, address- ing limitations of existing hardware-dependent approaches and advancing social inclusion for the deaf community.

Index Terms—Automatic Speech Recognition (ASR), Indian Sign Language (ISL), Natural Language Processing (NLP), As- sistive Communication Technology, GIF-Based Gesture Mapping, Real-Time Translation, Flask Web Application.