Gives Claude five web operations without requiring API keys: extract structured content (titles, headings, links, images, metadata) from URLs, search via DuckDuckGo, monitor site uptime and response times, convert HTML to markdown, and batch extract from multiple URLs in parallel. Reaches for this when you need web scraping or search capabilities in your agent workflows but don't want to manage authentication or run a headless browser. Built on direct HTTP requests with 15 second timeouts and strips out scripts and ads for clean extraction. Works through npx so there's no global installation step.
claude mcp add --transport stdio io.github.rog0x-web -- npx -y @rog0x/mcp-web-tools