MCP Server for PDF Q&A

A Model Context Protocol (MCP) server that lets LLMs answer questions over large PDF document corpora. Documents flow through an automated ETL pipeline — extraction, chunking, then embedding — before being indexed in a vector database for similarity-based retrieval.

Retrieval and orchestration are built with LangChain, and the conversational interface supports multi-turn Q&A with context-aware responses, so follow-up questions resolve correctly against earlier answers and retrieved passages.