Scrapling MCP: Giving Claude Code a Real Web Scraper

Short summary

The author integrates Scrapling, an open-source adaptive web-scraping framework, with Claude Code via its built-in MCP server, giving the AI agent ten tools for fetching, rendering, and stealth-scraping web pages including bypassing Cloudflare. The setup involves cloning the repo, installing into an isolated Python 3.14 venv, and registering the MCP server globally in Claude Code's config file. Key gotchas include backing up the config before editing, restarting Claude Code to load the new server, and using a dedicated venv to isolate dependency risks.

• •Scrapling's bundled MCP server gives Claude Code ten web-scraping tools including stealth fetches that bypass Cloudflare
• •Installation requires a Python venv, Playwright browser binaries (~700MB), and a one-block JSON config edit in ~/.claude.json
• •Three gotchas: back up config first, restart Claude Code to load MCP servers, and isolate dependencies in a dedicated venv

Generated with AI, which can make mistakes.