Parse web URLs into clean Markdown content by removing page clutter.
npx skills add https://github.com/tyler-dot-earth/.dotfiles --skill defuddle-tyler-dot-earth
Please help me install this Agent Skill. Skill: defuddle Source: https://github.com/tyler-dot-earth/.dotfiles/tree/main/agents/skills/defuddle Command: npx skills add https://github.com/tyler-dot-earth/.dotfiles --skill defuddle-tyler-dot-earth
Defuddle eliminates clutter and navigation from web pages, delivering clean, readable Markdown content to save tokens during reading or analysis.
Install Defuddle if needed and run defuddle parse <url> --md to generate clean Markdown.
High-intent search queries and answers about installing and using this skill.
To extract clean Markdown from a web page, the Defuddle tool removes ads, sidebars, and navigation boilerplate to isolate core article text. This significantly reduces token usage when processing online documentation, articles, or blogs for analysis.
The best way to remove web page clutter is using a content extraction tool that strips navigation and ads, outputting only the core text. This delivers clean, readable content formatted as Markdown by default to minimize token consumption.
Yes, extracted web content can be output as HTML or JSON instead of Markdown. While Markdown is the default format for token savings, optional flags allow you to generate HTML or JSON to support downstream processing workflows.
You need to install the Defuddle tool itself to parse web URLs into Markdown, but it has no external dependencies. Once installed, you can run a parse command with your target URL to generate clean, token-friendly content.
Content extraction works effectively on online documentation, articles, blogs, and standard web pages. It processes noisy URLs to remove clutter, producing clean Markdown formatted text suitable for faster summarization and reduced token usage.
WizarckExtract clean markdown from web pages via the Defuddle CLI.
lolwierdConvert URLs and local documents into Markdown using uvx markitdown.
XJTLUmediaExtract structured content and metadata from public web URLs.
sungaoxiang-backendExtract clean text from WeChat public account articles via Playwright.
