MCP server that wraps wget to download complete websites and adjust links for local browsing.
https://github.com/pskill9/website-downloaderStop copying and pasting HTML when you need to work with website content offline. This MCP server gives your AI assistant the ability to download complete websites with all assets intact, ready for local browsing and development.
You know the drill: find a useful site, try to save it locally, watch half the assets break, spend time fixing relative paths. This server handles the entire process—from initial crawl to link conversion—so you get a fully functional offline copy every time.
Built on wget's proven website mirroring capabilities, it downloads recursively while preserving the complete site structure. CSS, images, JavaScript, and internal links all work exactly as they do online.
Complete Site Preservation: Downloads all page requisites automatically—no missing stylesheets or broken image references
Smart Link Conversion: Transforms absolute URLs to relative paths so everything works offline without modification
Controlled Crawling: Set depth limits to grab just what you need, from single pages to entire site architectures
Domain Scoping: Automatically restricts downloads to the target domain, preventing runaway crawls
Local Development Ready: Generated sites work immediately in any local server or file browser
Research and Documentation: Archive technical documentation, tutorials, or reference sites before they change or disappear. Perfect for maintaining stable references during long projects.
Competitive Analysis: Download competitor sites for offline analysis without repeated requests. Study page structures, asset organization, and implementation patterns at your own pace.
Development Templates: Grab complete examples of interesting layouts or interactive components for local experimentation and learning.
Client Site Backups: Create offline copies of client websites for troubleshooting, testing, or archival purposes.
Network-Limited Development: Download resources when you have good connectivity, work offline later without dependency on external sites.
Add it to your MCP configuration and start downloading through natural conversation:
{
"mcpServers": {
"website-downloader": {
"command": "node",
"args": ["/path/to/website-downloader/build/index.js"]
}
}
}
Then simply ask your AI assistant: "Download the complete TailwindCSS documentation site for offline reference" or "Archive this competitor's landing page with all assets intact."
The server handles wget installation validation, directory management, and progress tracking. You focus on what you need downloaded, not how to make it work.
Whether you're building a resource library, preparing for offline development, or just want reliable local copies of useful sites, this MCP server eliminates the friction between "I need this site offline" and actually having it work properly on your machine.