website-downloader

MCP.Pizza Chef: pskill9

A single-purpose helper: give it a web address and it copies the pages, images, and styling into a folder on your machine, rewriting the links so the copy works with no internet connection. No account or key is needed, but the wget program must already be installed and you have to build the code yourself first. The project has had no code changes since December 2024. Note too that the address you supply is passed straight to your computer's command line without being cleaned, so only point it at addresses you trust.

Unmaintained · no code changes since December 2024
Files/PDF
Web/Research

Use This MCP server To

Grab a documentation site to read on a plane Archive a blog before the author takes it down Keep an offline snapshot of a product manual Download one page plus its images without following links Copy a recipe site into a folder I can browse anytime

README

Website Downloader MCP Server

This MCP server provides a tool to download entire websites using wget. It preserves the website structure and converts links to work locally.

Google Workspace Server MCP server

Prerequisites

The server requires wget to be installed on your system.

Installing wget

macOS

Using Homebrew:

brew install wget
Linux (Debian/Ubuntu)
sudo apt-get update
sudo apt-get install wget
Linux (Red Hat/Fedora)
sudo dnf install wget
Windows
  1. Using Chocolatey:
choco install wget
  1. Or download the binary from: https://eternallybored.org/misc/wget/
    • Download the latest wget.exe
    • Place it in a directory that's in your PATH (e.g., C:\Windows\System32)

Usage

The server provides a tool called download_website with the following parameters:

  • url (required): The URL of the website to download
  • outputPath (optional): The directory where the website should be downloaded. Defaults to the current directory.
  • depth (optional): Maximum depth level for recursive downloading. Defaults to infinite. Set to 0 for just the specified page, 1 for direct links, etc.

Example

{
  "url": "https://example.com",
  "outputPath": "/path/to/output",
  "depth": 2  // Optional: Download up to 2 levels deep
}

Features

The website downloader:

  • Downloads recursively with infinite depth
  • Includes all page requisites (CSS, images, etc.)
  • Converts links to work locally
  • Adds appropriate extensions to files
  • Restricts downloads to the same domain
  • Preserves the website structure

Installation

  1. Build the server:
npm install
npm run build
  1. Add to MCP settings:
{
  "mcpServers": {
    "website-downloader": {
      "command": "node",
      "args": ["/path/to/website-downloader/build/index.js"]
    }
  }
}

website-downloader FAQ

Do I need an account or a paid key?
No. There is nothing to sign up for and no key to paste. You do need the free wget program installed on your computer, which the setup instructions walk through for Mac, Windows, and Linux.
Is this project still being worked on?
It has had no code changes since December 2024, so treat it as finished rather than actively maintained. It is small enough that it still works, and the author has not named a replacement.
How hard is setup?
This one is for confident users. The code is not published as a ready-made package, so you download it, run a build step, and then point your assistant's settings at the built file on disk.
Can I use this to save a documentation site for offline reading?
Yes. Give it the site's address and a folder, and it pulls the pages along with the images and styling, then rewrites the links so everything opens locally.
What can it change on my computer?
It writes new files into whatever folder you name, or your current folder if you name none. It does not read or send your existing files, and it does not delete anything.
Is there anything I should be careful about?
Two things. The web address and folder you give it are handed to your computer's command line without any cleaning, so a booby-trapped address could run other commands; only use addresses you trust. And unless you set a depth limit, it keeps following links until the entire site is copied, which can take a long time and use a lot of disk space.
Which apps does it work in?
Any app that supports MCP servers you configure by hand, including Claude Desktop, Cursor, and similar tools. There is no hosted version to click into.