Documents

defuddle-web-cleaner

Extract clean article text, metadata, and Markdown from URLs, HTML, or web page text.

What it does

Extract readable article content from a URL, raw HTML, or web page text. The parser removes navigation, sidebars, ads, comments, footers, and social buttons, then returns title, author, site, publication date, and Markdown content by default. A JSON option includes description, cleaned content, and Markdown.

When to use it

  • Saving articles to Obsidian
  • Building research datasets
  • Preparing webpages for LLM processing
  • Converting web pages to Markdown

The skill document

name: defuddle-web-cleaner description: extract clean article content from web pages using defuddle. use when a user provides a url or html and wants the readable article text, markdown version, or structured metadata. helpful for web scraping, research workflows, note taking, obsidian clipping, and converting web pages to markdown.

Defuddle Web Cleaner

Extract the main readable content from a web page.

This skill removes unnecessary elements such as:

  • navigation bars
  • sidebars
  • ads
  • comments
  • footers
  • social buttons

The result is clean article content.

Supported Inputs

  1. URL
  2. Raw HTML
  3. Web page text

Output Format

Default output:

Title
Author
Site
Published date

Markdown article content

Alternative output (JSON):

{ title, author, site, description, published, content, contentMarkdown }

Processing Steps

  1. Detect input type
  2. Load page HTML
  3. Run Defuddle parser
  4. Extract metadata
  5. Convert to Markdown if requested
  6. Return clean content

Example

Input:

https://example.com/blog/ai

Output:

Title: AI is Changing Everything
Author: Jane Smith
Site: Example Blog

Markdown:

AI is Changing Everything

Artificial intelligence is transforming industries...

Tips

Use this skill when:

  • saving articles to Obsidian
  • building research datasets
  • cleaning webpages for LLM processing
  • summarizing articles

Questions people ask

What input formats can I provide?
Provide a URL, raw HTML, or web page text. The workflow detects the input type and loads page HTML before parsing.
Which page elements are removed?
It removes unnecessary elements such as navigation bars, sidebars, ads, comments, footers, and social buttons to isolate the main readable content.
What output formats are available?
The default output contains title, author, site, published date, and the article in Markdown. JSON output can include title, author, site, description, published, content, and contentMarkdown.

Related skills

Git

Handle Git changes, conflicts, history recovery, and collaboration with repository-aware safety checks.

527 installs31 stars

OpenClaw Token Optimizer

Reduce OpenClaw token waste with lazy context, model routing, heartbeat planning, and local budget checks.

459 installs30 stars

Developer

Turns development work into scoped changes, verified tests, reviewable PRs, and reversible releases.

179 installs11 stars

Google Ads

Query Google Ads accounts and performance data with GAQL through managed OAuth.

261 installs20 stars

Vue

Build, debug, and review Vue 3 code across reactivity, components, state, routing, forms, and performance.

121 installs8 stars