With an emphasis on the Mahabharata, a complex narrative with many characters, events, stories, and deep relationships, this paper investigates character-centric summarization and keyphrase extraction in Indian mythology. Through refinement of transformer-based models such as BART, PEGASUS, and T5 and using the dataset, \(DS_{character}\) , the study generates character-specific summaries. In the dataset, the character-centric sentences emphasize the roles, actions, and significance of individual characters in the text. In case of capturing the coherence and consistency of character-specific summaries when the models were assessed using ROUGE and BLEU scores, PEGASUS performed the well overall. This study also applied KeyBERT for the extraction of character-specific keyphrases which achieves high recall and precision in recognizing important themes linked to characters. To understand the significance of each character in the story, later on we focused on to implement the visual representation of the character-specific summaries and keyphrases. This research offers a detailed understanding of how character-centric mythological sentences are processed and makes recommendations for future research areas to enhance model performance in character-centric tasks.