CONTRACTSENSE AI: AN INTELLIGENT LEGAL DOCUMENT ANALYSIS SYSTEM USING RETRIEVAL-AUGMENTED GENERATION

ContractSense AI is a production-ready Retrieval-Augmented Generation (RAG) web application that analyzes
legal contracts for risk, surfaces dangerous clauses, and enables interactive question-answering over uploaded
PDF documents. The system combines FastAPI with LangChain orchestration, a locally running sentencetransformer embedding model (all-MiniLM-L6-v2), and the Groq-hosted Llama 3.3 70B large language model
for fast inference. A custom NumpyVectorStore provides lightweight vector similarity search without external
database dependencies. The frontend features an animated Three.js landing page and a fully interactive threepanel dashboard with risk gauge, red-flag cards, TL;DR summary, and RAG-powered conversational chat. All
embedding computation runs locally — no cloud embedding API is required — making the system inherently
privacy-preserving and cost-free to operate at the inference layer. The system is designed for deployment on
standard personal computing hardware without requiring GPU acceleration, making it accessible to students,
freelancers, and small organisations with limited IT infrastructure