Exposes two tools for web scraping workflows: html_to_md converts HTML into clean Markdown, and extract_text strips straight to plain text. Built for agents that need to ingest web content without the noise of tags, scripts, and styling. Part of a fourteen-server collection (mcp-stack) that handles transforms LLMs routinely hallucinate. Ships as an npm package you can run via npx in Claude Desktop, Cursor, Cline, or any MCP client. Install it when your agent is fetching pages and you want deterministic parsing instead of letting the model imagine what the HTML contained or guess at Markdown conversion rules.
claude mcp add --transport stdio io.github.mukundakatta-html-to-markdown-mcp uvx html-to-markdown-mcp