Published Paper Details:

It is difficult for students, researchers, legal professionals, and citizens to extract clear and useful information from Indian legal documents due to their large, intricate, and linguistically dense structure, which includes constitutional provisions, statutes, and judgment texts. This study proposes a Retrieval-Augmented Generation (RAG) framework for comprehending and summarising Indian legal texts to overcome these challenges. The framework is intended to produce precise, context-aware, and legally sound results. In order to minimise hallucinations and guarantee that generated responses stay true to reliable legal sources, the system uses document segmentation, chunking, domain-specific embeddings, and vector-based semantic search. The framework facilitates both brief and detailed point-wise summarisation, contextual question answering, and automated extraction of important legal insights, in addition to generating succinct and context-rich summaries. The comprehension and accessibility of complex legal material are further improved by additional user-oriented features like embedded YouTube explanations and PDF downloads. When compared to standalone generative models, experimental evaluation shows notable gains in contextual precision, faithfulness, and relevance. All things considered, this RAG-based framework contributes to the development of reliable AI for the legal field by offering a scalable, dependable, and user-friendly solution for navigating, interpreting, and learning Indian legal texts.