Read a web page as clean LLM-ready markdown (main content, no nav or scripts) for research, RAG and summarization. Respects robots.txt.