Extract clean markdown from web pages using the Defuddle CLI.
npx skills add https://github.com/allemaar/open-skills --skill defuddle-allemaar
Please help me install this Agent Skill. Skill: defuddle Source: https://github.com/allemaar/open-skills/tree/main/skills/defuddle Command: npx skills add https://github.com/allemaar/open-skills --skill defuddle-allemaar
defuddle parse <url> --md to get a token-efficient markdown version ready for summarization or saving as a note. ## Quick Start Ask the agent to read and summarize the article at a given URL using defuddle instead of a standard web fetch.High-intent search queries and answers about installing and using this skill.
Run defuddle parse <url> --md to convert a web page into markdown with navigation, ads, and clutter removed. Add -o content.md to save the result directly to a file.
Use Defuddle for standard articles, blog posts, and documentation pages where you want readable content without clutter and lower token usage. Skip it for APIs, JSON endpoints, or raw file downloads.
Install it globally with npm by running npm install -g defuddle. Once installed, the defuddle command is available for parsing any URL.
Yes, use the -p flag with a property name, such as defuddle parse <url> -p title, -p description, or -p domain, to extract a single metadata field instead of the full content.
Defuddle supports markdown via --md, JSON containing both HTML and markdown via --json, plain HTML with no flag, and individual metadata properties via -p <name>.
drewid74Archive websites into a searchable personal knowledge base with full-text search.
devCharuzuIngest sources into Obsidian and update entity pages with cross-references.
SigmacodeatCoordinate multiple external research tools for broader information gathering.
eddmannRender dynamic web pages with Playwright to capture HTML, text, and screenshots.