Scrape Website Content is a WordPress ability from Page Builder: Pagelayer – Drag and Drop website builder — usable via Easy MCP AI, which connects Claude, ChatGPT, and any AI assistant to your site.

Install Free Plugin →

Scrape Website Content

Read-only

Fetch a live URL and extract its raw title, meta description, headings, paragraph text, and image URLs for reference material.

Requires the Pagelayer plugin on WordPress 6.9+ (or another source of wp_register_ability()) for the Abilities API — registration is skipped entirely via a function_exists() guard when it isn't available, and ability categories similarly require wp_register_ability_category(). Most abilities require the 'edit_posts' capability; update-global-styles, create-menu, and delete-menu require 'manage_options'. search-images additionally needs a Pexels API key configured on the Pagelayer AI Agents settings page, and get-library-sections/import-library-section call Pagelayer's own hosted library API. Four near-duplicate abilities (get-widget, get-widgets-summary, edit-layout, navigator) are registered in source with mcp_public:false — fully callable over the REST Abilities API but intentionally left out of the advertised MCP tool list as duplicates of get-widget-schema, list-widgets, update-page, and get-page respectively — and are excluded from this listing.

What it does

Performs a server-side HTTP GET of the given URL and regex-extracts the page title, meta description, up to 20 image URLs (resolved to absolute), and the h1/h2/h3 headings plus the first 12 body paragraphs. It performs no layout or design decisions and creates nothing — it exists purely to hand an AI client real source material, so it doesn't fabricate facts like pricing or testimonials, before it designs an original page with create-page or create-website.

See it in action

You ask your AI assistant

"Scrape https://example-competitor.com and pull out its headings and image URLs so I can reference them while building a similar page."

Inputs

Name Type Required Description
url string Yes Target website URL to read (e.g. https://example.com/).

Returns

An object with source_url, title, meta_description, headings (h1/h2/h3 arrays), paragraphs (up to 12), images (up to 20 URLs), and a note reminding the caller not to fabricate content not actually found.

Permission

Requires the 'edit_posts' capability.

Newsletter

The AI + WordPress space moves fast. Keep up.

New tools, workflow ideas, and product updates — be the first to know what's next.

No spam, unsubscribe anytime.