AI-Driven Legal Assistance Using Rag

Abstract

This project presents an AI-powered legal assistant designed to automate the drafting, interpretation, and retrieval of legal documents. The system leverages a Modified Retrieval-Augmented Generation (MRAG) architecture, where uploaded legal files are processed using LangChain, split into meaningful text chunks, and converted into vector embeddings stored in ChromaDB for efficient semantic search. When a user asks a legal query or requests document drafting, the system retrieves the most relevant content and generates accurate, context-aware responses through an LLM-based reasoning model.

The platform supports essential legal functions including clause extraction, legal Q&A, contract drafting, advocate search, and automated email communication. It provides a secure user workflow with authentication, private document processing, and instant output delivery through an intuitive web interface. By reducing manual effort, minimizing human error, and offering fast, reliable legal guidance, the system significantly improves accessibility to legal assistance. This makes it suitable for individuals, small organizations, and environments where quick, cost-effective, and accurate legal decision-making is required.