Skim MCP server

Skim is a helper that lets your AI read web pages properly. Instead of messy HTML, you get clean text with the title, author, date, and language at the top. It is handy for anyone who asks their AI to summarize articles, compare pages, or pull facts from a site.

What this MCP server does

You ask your AI to read a web page, a few pages, a PDF, or a whole small site. The AI passes the address to this helper. The helper sends it to Skim, which fetches the page and strips out ads, menus, and clutter. Skim sends back tidy Markdown with a small header showing the title, author, date, and language. Your AI then reads that and answers you.

Click to zoom

What you can do with it

• Read a single web page and summarize it
• Read up to 10 pages in one go and compare them
• Pull specific fields like product name, price, or job details as structured data
• Crawl a small site and list its page titles
• Read a public PDF and give you the outline
• Watch pages for changes and check what changed later
• Poll a Skim Signal feed for the latest items in a topic

Try asking your AI

• “Read https://en.wikipedia.org/wiki/HTTP_402 and summarize it.”
• “Read these three pricing pages and compare them: https://a.example/pricing https://b.example/pricing https://c.example/pricing”
• “Extract the product name, price, and availability from https://example.com/products/notebook as JSON.”
• “Crawl https://example.com (max 10 pages) and list the page titles.”

What it gives back to you

You get clean Markdown text in the chat, with a short header at the top showing title, author, date, and language. For extraction, you get structured fields like name, price, and availability. For crawling, you get one clean page per result. For watches, you get a list of what changed.

Before you start

Who it's for

Anyone who asks their AI to read, summarize, or compare web pages, like researchers, analysts, writers, and office workers doing desk research.