Kishan Sai Saguturu, Bhuiyan Md Moinuddin, Abinav Satya Sripathi, Lokesh Umamaheswari Ethirajan, Venkata sai kumar Erla and Tirth Chheta, University of New Haven, USA

The rapid growth of online information has made automatic text summarization essential for productivity and knowledge access. While transformer-based abstractive summarizers dominate current research, they often demand high computational resources and reduce interpretability. This work presents a lightweight extractive summarization framework with a web-based interface that balances efficiency, scalability, and usability. The system accepts both user-provided text and live news articles, applying frequency- and position-weighted ranking to identify key sentences. Built with Python libraries such as spaCy, BeautifulSoup, and Streamlit, it generates concise summaries with low latency and supports real-time interaction. Experimental evaluation with ROUGE metrics and tokenization accuracy confirms the frameworks reliability and effectiveness. Designed for extensibility, the system outlines future directions including neural abstractive integration, multilingual support, and sentiment-aware summarization. By bringing together the interpretability of foundational methods and pathways to modern advancements, this work provides a practical bridge between academic study and real-world applications

NLP, Extractive Framework, BeautifulSoup, spaCy, Text summarization