Why use Defuddle instead of standard web fetching?

Defuddle removes 'noise' like headers, footers, and ads, which reduces the number of tokens Claude needs to process, making analysis more accurate and cost-effective.

Can I save the extracted content to a file?

Absolutely. You can use the --md -o flags to save the cleaned Markdown directly to your project workspace for later use.

Does it support formats other than Markdown?

Yes, while Markdown is the preferred choice for AI context, Defuddle also supports JSON output and specific metadata extraction like titles and domains.

Is this skill ideal for technical documentation?

Yes, it is highly effective for technical documentation as it isolates the primary instructional content from complex site-wide navigation systems.

Do I need to install anything for this skill to work?

Yes, this skill requires the defuddle-cli package, which can be installed globally via npm using 'npm install -g defuddle-cli'.

Defuddle Web Content Extractor

Name: Defuddle Web Content Extractor
Author: kepano

bykepano

•

11,747

•

ウェブスクレイピングとデータ収集

Extracts clean, distraction-free Markdown content from web pages to optimize token usage and context window efficiency.

Defuddle is a specialized skill designed to streamline web content ingestion for Claude. By leveraging the Defuddle CLI, it strips away navigation menus, advertisements, and clutter from URLs, providing a lean Markdown version of articles, blog posts, and documentation. This approach significantly reduces token consumption while ensuring the agent receives only the most relevant information for high-quality analysis or summarization tasks.

主な機能

01Supports metadata extraction including titles and descriptions

02Optimizes token usage for long-form web content

0311,747 GitHub stars

04Automatically removes navigation, ads, and boilerplate clutter

05Converts complex HTML pages into clean, readable Markdown

06Saves output directly to files for workspace integration

ユースケース

01Analyzing online documentation and technical blog posts without sidebar noise

02Gathering clean web data for knowledge base creation and management

03Summarizing news articles and research papers for AI context

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add kepano/obsidian-skills defuddle

For use in Claude.ai and ChatGPT

Download Skill