Text summarization is a significant undertaking in natural language processing (NLP) that plans to consolidate enormous volumes of messages into more limited, significant rundowns while saving the fundamental data. Customary rundown strategies, basically extractive in nature, frequently neglect to deliver reasonable and relevantly precise synopses as they just select sentences without rewording or deciphering the substance. Late headways in brain networks have altered this field, especially through the advancement of models that consolidate extractive and abstractive procedures. This paper proposes an original Text Summarization structure that uses a cross-over approach, incorporating BiLSTM networks for extractive synopsis with a transformer-based model for abstractive summarization. The proposed system means to improve the nature of synopses by guaranteeing both intelligibility and significance. Broad tests directed on different datasets exhibit that the proposed system essentially beats customary techniques, as well as cutting-edge brain rundown models, across different measurements such as ROUGE, BLEU, and METEOR. The system’s versatility across various areas and its capacity to deal with complex texts feature its true capacity for genuine applications, including legitimate, clinical, and scholarly report rundown. Future work will zero in on improving the system for better computational effectiveness and investigating its application in other NLP assignments.