Scrape Any Website
Your agent scrapes what web fetch can't: stealth browsing clears Cloudflare, spiders run full crawls, pages come back as clean Markdown. The official skill from the Scrapling author.
Install the skill or give me an instruction how to install it - scrapling-official from https://github.com/D4Vinci/Scrapling Free Open source Works in Claude, Cursor, Codex & more
What it does
Ask your agent to scrape a page and it works, including the pages that normally fail: JavaScript-heavy apps, sites behind Cloudflare, pages the built-in web fetch cannot reach. scrapling-official is the official Scrapling skill, written by the library author, and it triggers on its own whenever you ask to scrape, crawl, or extract data.
Under the hood the agent runs Scrapling's extract commands, escalating from a plain HTTP request to JS rendering to a stealth browser only as needed, and saves the result as clean Markdown, text, or HTML. Every scrape uses the ai-targeted flag, which keeps only the main content and strips hidden elements that could carry prompt injection.
When you need more than one page, the skill's bundled docs teach the agent to write full Python scrapers and Scrapy-like spiders, no web searches required.
When to use it
- Your agent's built-in web fetch just failed
The skill steps up from plain HTTP to JS rendering to a stealth fetch and saves the page straight to Markdown.
- The data you need sits behind Cloudflare
The stealth fetcher gets past Turnstile and interstitial checks without a captcha service or an API key.
- You want the scraper written for you
The agent writes working Scrapling code from its bundled docs, a quick fetch or a full spider, with no web searches needed.
- Scraped pages are eating your context window
Every CLI scrape pulls just the main content, blocks ads to save tokens, and strips hidden elements that could carry prompt injection.
FAQ
Which coding agents does it work with?
scrapling-official installs into coding agents such as Claude Code. It is the official skill, written by the library author and shipped inside the Scrapling repo, and everything runs locally on your machine.
Do I need API keys or a captcha-solving service?
No. scrapling-official gets past Cloudflare Turnstile and Interstitial pages through browser automation alone, so there is no solver service to pay for and no credentials or API keys to set up. Proxies and CDP endpoints are optional extras, not requirements.
Is it safe to let an agent read untrusted pages?
scrapling-official guards against prompt injection by forcing the --ai-targeted flag on every CLI scrape, which extracts only the main content and sanitizes hidden page elements before they reach your agent. It also turns on ad blocking automatically to save tokens.
Do I need a slash command to run it?
No. scrapling-official triggers on its own when you ask your agent to scrape, crawl, or extract data, when the built-in web fetch fails, when a site has anti-bot protection, or when you ask for Python scraping code.