Skip to content
← All projects
GenAI / RAGIn progress

Indian Legal AI Assistant - RAG over Indian Law

A Streamlit RAG assistant that answers plain-English Indian legal questions from a corpus of legal PDFs and cites source documents and pages in the response.

Overview

A legal-domain RAG assistant that answers questions from Indian law PDFs with source citations.

Promoted as a live demo only after retrieval quality, citation behavior, and safety wording are cleaned up.

Impact

  • Built a LangChain pipeline over 39 PDFs, 2,390 pages, and 9,851 chunks using bge-small embeddings.
  • Refactored the app into modular loader, embedding, vector-store, LLM, and config layers.
  • Remaining work: retrieval filtering and chat-history fixes, index persistence, and deployment.

What it shows

  • Document loading, chunking, embeddings, vector search, and citation-oriented generation.
  • Domain adaptation for a practical high-context retrieval problem.
  • Modular RAG app structure across loaders, embeddings, vector store, LLM, and UI.

Demo plan

  • Deploy as a Hugging Face Space or Streamlit Community Cloud app once retrieval filtering is fixed.
  • Show citation snippets and a legal disclaimer clearly before sharing publicly.

Stack