In today’s globalised and linked world, the need for multi-modal translation systems has increased and it requires innovative solutions to address communication gaps. This research work investigates the evolutions of AI-driven natural language processing (NLP) systems designed for multi-modal translation applications, including spoken and written languages, as well as sign language. The suggested approach uses deep learning models such as transformers, with advanced computer vision methods to provide seamless translations across several media, including text, audio, and sign language. The system utilises video input to proficiently recognise sign language motions and translate them into spoken or written language, and conversely. Training the NLP models on broad multilingual datasets guarantees great translation accuracy across several languages. The proposed model emphasises real-time performance, enabling users to participate in seamless, natural discussions across language boundaries, thereby improving communication efficiency. This advanced AI-based translation system seeks to bridge the significant gap between spoken and sign languages, enhancing inclusion and accessibility for deaf and hard-of-hearing individuals. The initiative improves user experience by overcoming communication obstacles and provides successful cross-lingual interactions across several modalities, fostering a more inclusive and communicative society.