Scraper

io.github.ofershapPublished by GitHub account ofershapv1.0.1Updated Oct 24, 2026

Web scraping MCP — extract clean markdown, links, and metadata from any URL.

VerifiedSTDIODesktop onlyWeb Search & ScrapingKnowledge & Memory

Overview

AI-generated overview

A local MCP server that fetches public web pages and returns clean markdown, links, and metadata for an AI assistant to read and summarize.

What you can do
Fetch a URL and extract clean, readable text as markdown using Mozilla Readability parsing in Node, without a headless browser. Pull all links from a page with their href and anchor text. Read page metadata such as title, description, OG tags, canonical URL, and favicon. Search within a page for a query string and get the matching lines. Batch-scrape several URLs at once and get a title and excerpt for each.
When to use it
Useful when an assistant needs to read a documentation page, blog post, article, or API reference and summarize or quote it, instead of the user copy-pasting content. It suits server-rendered pages; for JavaScript-heavy single-page apps a browser automation server is a better fit.
Requirements
Runs locally as a stdio process started with npx from the npm package mcp-server-scraper, so Node.js and network access are needed. No accounts, API keys, environment variables, or headers are required.
Before you install
It fetches whatever URLs the assistant is given and sends those requests to third-party sites, so avoid pointing it at internal or private addresses. It only reads pages and does not log in, submit forms, or change remote data. It does not execute JavaScript, so client-rendered pages may return little or no content.

Installation

In SourceWeft

  1. Open Scraper in the dashboard and add it to a workspace.
  2. Enable the server for the chats that should use its tools.

Desktop only via STDIO. STDIO servers start a local process, so they need the SourceWeft desktop host.

Other MCP clients

Follow the launch instructions in the repository.

README

mcp-server-scraper

[npm version] [npm downloads] [CI] [TypeScript] [License: MIT] [Agent Plugins]

Extract clean, readable content from any URL. Returns markdown text, links, and metadata. No API keys, no config. A free alternative to Firecrawl for scraping docs, blogs, and articles.

bash
npx mcp-server-scraper

Works with Claude Desktop, Cursor, VS Code Copilot, and any MCP client. No accounts or API keys needed.

[Install in Cursor]   [Add to VS Code]

[MCP server for web scraping, content extraction, and URL metadata]

Demo built with remotion-readme-kit

Why

When you're working with an AI assistant and need to reference a docs page, a blog post, or an API reference, you usually end up copy-pasting content manually. Tools like Firecrawl solve this but require a paid API key. This server does the same thing for free. It fetches a URL, runs it through Mozilla Readability (the same engine behind Firefox Reader View), and returns clean markdown. It works well for server-rendered content like documentation sites, blog posts, and articles. It won't handle JavaScript-heavy SPAs, but for the most common use case of "read this docs page and summarize it," it does the job.

Tools

ToolWhat it does
scrape_urlExtract clean text content from a URL (Readability-powered)
extract_linksGet all links with href and anchor text
extract_metadataGet title, description, OG tags, canonical, favicon
search_pageSearch for a query string within the page, return matching lines
scrape_multipleBatch scrape multiple URLs, get title + excerpt per URL

Quick Start

Cursor

Add to .cursor/mcp.json:

json
{  "mcpServers": {    "scraper": {      "command": "npx",      "args": ["-y", "mcp-server-scraper"]    }  }}

Claude Desktop

Add to claude_desktop_config.json:

json
{  "mcpServers": {    "scraper": {      "command": "npx",      "args": ["-y", "mcp-server-scraper"]    }  }}

VS Code

Add to your MCP settings (e.g. .vscode/mcp.json):

json
{  "mcp": {    "servers": {      "scraper": {        "command": "npx",        "args": ["-y", "mcp-server-scraper"]      }    }  }}

Examples

  • "Scrape the API docs from https://docs.example.com and summarize them"
  • "Extract all links from this page"
  • "What's the OG image and description for this URL?"
  • "Search this page for mentions of 'authentication'"
  • "Scrape these 5 URLs and give me a summary of each"

How it works

Uses Mozilla Readability (the engine behind Firefox Reader View) plus linkedom for fast HTML parsing in Node. No headless browser needed. Works best with server-rendered pages: docs, blogs, articles, news sites.

Agent Plugins

This repo is an Agent Plugins 1.0.0 package: plugin.json, portable mcp.json, and skills/ ship together with the MCP server.

For Cursor, clone the repo and copy or symlink it to ~/.cursor/plugins/local/mcp-server-scraper, then reload the window. Skills and MCP show up under Customize > Plugins.

The Cursor and VS Code install buttons above still work: they add the same npx -y mcp-server-scraper stdio server as manual JSON.

FAQ

What is mcp-server-scraper?

A free MCP server that turns public web pages into clean markdown using Mozilla Readability. No Firecrawl or other scrape API key.

Does it run JavaScript or SPAs?

No. It fetches HTML and parses it in Node. Use a browser MCP for React dashboards and other client-rendered sites.

How is this different from Firecrawl?

Firecrawl is a hosted scrape API with billing. This server runs locally via npx, costs nothing, and fits doc/blog/article URLs.

Can I install it as an Agent Plugin in Cursor?

Yes. Use the local plugin path under ~/.cursor/plugins/local/mcp-server-scraper so the bundled web-scraping skill loads with the MCP config.

Do I need API keys or env vars?

No. Point your MCP client at npx -y mcp-server-scraper only.

Development

bash
npm installnpm run typechecknpm run buildnpm test

See also

More MCP servers and developer tools on my portfolio.

Author

[Made by ofershap]

[LinkedIn] [GitHub]

README built with README Builder

License

MIT © Ofer Shapira

Source: README.md at commit 7602c29

Tools

0
SourceWeft has not started this package in its sandbox yet, so its tools are not known.

Version history

1
  1. v1.0.1LatestSep 16, 2026

Score

50% of the applicable checks met

How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.

  • Reached in the last 30 days25 pts

    Keep the endpoint answering the MCP handshake; SourceWeft probes it every 30 days.

  • Auth declared10 pts

    Declare the credential headers in the registry entry, or keep the endpoint answering the probe so its auth can be classified.

  • Tools listed15 pts

    Answer tools/list without a credential, or ship a package that starts in a sandbox without one.

  • DescriptionRequired15 pts
  • Logo10 pts
  • Documentation15 pts
  • License10 pts