• RESEARCH
  • CONTENT
  • AUDITS
  • HUMAN

How to extract website metadata with your AI

Connect Microlink once, then let your agent scrape site details from any link you provide.

Ask AI about this page

Short answer

How do I pull website info using my AI?

Give your AI a URL, and it uses Microlink to grab the page title, description, and social media images. You get structured data ready for your research or content tasks.

Where the data lands.

Organized site information.

The AI turns raw URLs into structured data points.

RESEARCH

Clean text data

The AI pulls titles and descriptions from the page. You get clean text for your notes or databases.

CONTENT

Social media assets

Your agent finds the specific image URLs used for social previews. This helps you prep visual content.

AUDITS

SEO snapshots

The AI checks if a page has the right tags. You can see exactly what a crawler sees.

HUMAN

Manual selection

The AI hands you the extracted metadata. You decide which specific fields to use in your final work.

The workflow

What your AI does when a URL arrives.

The AI handles the heavy lifting of parsing web pages.

  1. Fetch page data

    Your agent sends the URL to the connector to pull the underlying metadata.

    extract_metadata

  2. Parse tags

    The AI identifies the specific title, description, and Open Graph image tags.

    extract_metadata

  3. Format output

    The agent organizes the scraped info into a clean list for you.

    extract_metadata

  4. Deliver results

    The information is presented directly in your chat interface.

    extract_metadata

Try it

Copy these to start.

Use these prompts to trigger the extraction.

Starting points

These are opening lines. Swap the URLs for the ones you actually want to scrape.

Microlink Connector

You're all set. Choose your MCP client and follow the setup instructions.

Connector linkhttps://edge.vinkius.com/vk_preview_oQj3CwHyFt2PqndYyKroiBBKDimeR5ZnHQGVy4GR/mcp

Claude Desktop

Follow the steps below to connect in seconds.

  1. 1In Claude Desktop, open Settings → Connectors.
  2. 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
  3. 3Click Add and start a new chat — Microlink capabilities are ready to use.
Configuration · claude_desktop_config.jsonCopy
{
  "mcpServers": {
    "microlink-mcp": {
      "url": "https://edge.vinkius.com/vk_preview_oQj3CwHyFt2PqndYyKroiBBKDimeR5ZnHQGVy4GR/mcp"
    }
  }
}
Copy into chat04

  • Extract the title and description from https://example.com

  • Get the social media image URL for this link: https://example.com

  • Summarize the metadata for https://example.com and put it in a table

  • Check if https://example.com has a meta description

  • Claude
  • ChatGPT
  • Cursor
  • VS Code
  • Windsurf
  • Claude Code
  • JetBrains
  • Cline

Start here

Connect Microlink once, then ask.

One click connects the MCP to your client. Your credentials stay encrypted, and you can start scraping immediately.

Connect Microlink to your AI

FAQ

How this task behaves.

  • 01

    Can the AI change the website's metadata?

    No. The AI only reads and extracts existing data from the URL.

  • 02

    Does it work on any website?

    It works on any URL that has publicly accessible metadata tags.

  • 03

    Can the AI download the images for me?

    No. The AI extracts the image URLs, but you must download the files yourself.

  • 04

    Will it scrape content behind a login?

    No. The tool can only access data that is publicly visible on the web.

  • 05

    Can I use this to scrape entire articles?

    No. This specific tool is designed to pull metadata like titles and images, not full body text.

  • More questions about Microlink? The Connector page answers them. See everything the Microlink Connector can do

Connect Microlink to Claude, Cursor, ChatGPT & more