Awesome MCP › Browser Automation

cyberchitta/scrapling-fetch-mcp

⭐ 128 Python added to this list on 2025-05-14 repository created 2025-03-13

scrapling-fetch-mcp addresses a specific failure: an assistant asked to read a documentation page receives a bot-protection challenge instead of content. It uses Scrapling, which combines patchright — a stealth-patched browser automation driver — with curl-cffi, an HTTP client that reproduces browser TLS fingerprints, so requests resemble a real browser at both the network and the runtime level. Content is returned as clean HTML or converted markdown, ready for a model to read rather than as a page of protection-challenge markup. The README is explicit about intended scope: it is optimised for low-volume retrieval of documentation and reference material, and is not designed for high-volume scraping or data harvesting — a limit worth respecting both technically, since volume is what triggers escalated blocking, and in terms of what the tool is for. It requires Python 3.10 or newer with uv, is published on PyPI, and is Apache-2.0 licensed. Alongside the MCP server it ships a Claude Code skill, which the README recommends as the easiest path: installed into a project's .claude/skills directory, it makes Claude reach for this fetcher automatically when a URL turns out to be protected, with no explicit instruction from the user. The audience is developers whose agents need to read technical documentation on sites that block ordinary automated requests.

https://github.com/cyberchitta/scrapling-fetch-mcp

mcpweb-scrapingscraplingpatchrightpythonbot-protectionmarkdownfetch