defuddle

Name: defuddle
Availability: InStock
Author: antonyfmunoz

Extract clean markdown from web pages using the defuddle CLI.

1|Updated Mar 11, 2026

Search Tags:#metadata extraction #web content #cli tool #token reduction #markdown extraction #documentation ingestion #content cleaning

Ready to run

npx skills add https://github.com/antonyfmunoz/OS --skill defuddle-antonyfmunoz

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: defuddle
Source: https://github.com/antonyfmunoz/OS/tree/main/skills/tools/defuddle
Command: npx skills add https://github.com/antonyfmunoz/OS --skill defuddle-antonyfmunoz

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Defuddle removes the navigation, ads, and other web-page clutter that inflates tokens and makes LLM processing harder, replacing raw HTML with clean, readable markdown.

Core Features & Use Cases

Token-efficient markdown extraction: turns the main content of a standard web page into compact markdown using Readability-style parsing.
Metadata extraction: quickly pulls fields like title, description, and domain without needing to process the whole page content.
Practical fallback behavior: supports standard pages well and is intentionally not for .md URLs, authenticated pages, or JS-only sites.

Quick Start

Run defuddle parse https://example.com/docs/intro --md to get the page’s main content as clean markdown.

Frequently Asked Questions about defuddle

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I extract clean markdown from a web page for LLM workflows?▼

To extract clean markdown from a web page, use a CLI tool to parse the URL and remove navigation or ads. This reduces token waste by outputting only the readable main-body content as compact markdown for downstream processing.

What is the best way to reduce token waste when ingesting web content?▼

Reducing token waste during web content ingestion is achieved by applying Readability-style parsing to strip HTML clutter. This replaces noisy raw HTML with clean markdown, making the extracted documentation or articles efficient for LLM consumption.

Can I extract metadata like title and description without processing full web page content?▼

Yes, you can extract metadata like title and description without processing full web page content. The extraction mechanism quickly pulls specific metadata fields from a URL, avoiding the need to parse the entire main body.

Does web content markdown extraction work with JavaScript-only sites or authenticated pages?▼

Web content markdown extraction does not work with JavaScript-only sites or authenticated pages. The parsing mechanism is designed for standard pages and intentionally avoids .md URLs, providing practical fallback behavior for unsupported content.

How do I parse documentation pages into markdown using a CLI tool?▼

To parse documentation pages into markdown using a CLI tool, invoke the command with the target URL and the markdown flag. This immediately outputs the main content as clean markdown, ready for ingestion into your LLM workflow.