Extract tables precisely from PDFs and convert them to clean HTML for RAG pipelines, running fast on CPU without external dependencies.