Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{205414,
        author = {Makham Aravind and Dr.P Shyam Sunder},
        title = {LexiSummary AI: Advanced Legal Document Summarisation Framework Using Hybrid Extractive – Abstractive Transformer Models},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {13},
        number = {1},
        pages = {6627-6639},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=205414},
        abstract = {Currently, lawyers in the legal profession have more digital files than there is time available to analyze, including but not limited to: court decisions, legislative files, contracts, and case studies. These digital files are considerably larger than they were historically when professionals would analyze them using a pen, pencil, paper, book and so on. Analyzing legal files by hand has been a very time-consuming and tedious way of analyzing legal files for centuries; and lawyers have been very prone to making mistakes due to burnout or other reasons. Most of the previously developed systems to automate legal document analysis used extractive summaries to produce summaries; however, extractive summaries fail to accurately reflect the relationships between the content in the document and the overall concept of the document or the meaning contained in the document due to the lack of legal reasoning and context contained within the generated summary. The goal of this project is to develop a new web-based software application (Lexi Summary AI) to automate the summarizing of legal documents, and analyse the documents using state-of-the-art natural language processing (NLP) and transformer-based architectures (i.e. BART, T5, and current generative AI) methods, using a hybrid extractive/abstractive approach that allows the system to create summaries from legal documents that do not have a fixed length in tokens through an overlapping chunking strategy. Our system's architecture consists of: ingesting legal documents, preprocessing, semantically chunking, AI processing, and moving data between our cloud platform (Firebase) and a client. When we validated our recent prototype, we validated 15 test parameters on the quality of our summaries (e.g., error rate in parsing documents, number of noise documents removed, and percentage of the context of the summary preserved).},
        keywords = {Legal Document Summarization, Natural Language Processing, Transformer Models, , Deep Learning, Artificial Intelligence, LexiSummary AI.”},
        month = {June},
        }