scrapling-fetch-mcp is an MCP server designed to help AI assistants access text content from websites that implement bot-detection measures. It targets low-volume retrieval of documentation and reference materials (text/HTML only) and is not intended for broad data harvesting. The server exposes two Claude-friendly tools: Page fetching, which retrieves complete web pages with pagination to cover multi-page content, and Pattern extraction, which uses regular expressions to locate and extract specific content. The AI determines which tool to use based on the user's natural-language request, enabling seamless interaction without exposing scraping details. Protection modes provide three levels of bot-detection bypass: basic, stealth, and max-stealth, with progressively higher latency to accommodate tougher defenses. Setup typically involves Claude Desktop configuration and, for operation, the installation of browser binaries (a large initial download). The project is built on Scrapling for bot-detection bypass. Limitations include text-content-only output, possible authentication barriers, and variable performance depending on site complexity and protection level. The server aims to bridge the gap between browser-visible content and AI-accessible text.