The exponential growth of today’s modern and digital world demands an efficient multilingual text summarization, especially in linguistically diverse regions like India. This study presents a generative AI-based summarization web application tailored for majorly used Indian languages including Hindi, Bengali, Tamil, Telugu, and Marathi. By using transformer-based models such as BART [3], T5 [5] and mBART [17] from the Hugging FACE, our approach integrates adaptive token selection and semantic coherence optimization to produce fluent, accurate, and contextually relevant summaries. We have also fine-tuned these models on multilingual datasets like MKSum and XL-Sum [1, 10], addressing the challenges in preserving semantic depth across the languages. We also have evaluated our approach using BLEU, ROUGE, and human assessments. This work enhances the accessibility to multilingual content and thus, offering a scalable solution for the real-world applications in the diverse linguistic settings.