What does Defuddle do?

Defuddle CLI extracts clean, readable content from web pages by removing navigation, ads, and clutter, which reduces token usage compared to noisy page-fetch approaches.

How do I install Defuddle?

Run `npm install -g defuddle` to install Defuddle globally.

What output formats does Defuddle support?

Defuddle can output Markdown (default), JSON containing both HTML and markdown, raw HTML, or specific metadata properties like title, description, or domain.

How do I extract metadata like a page title with Defuddle?

Use the `-p` flag followed by the metadata property name, for example: `defuddle parse -p title`

Skill

Extract Clean Content from Web Pages

Name: Defuddle
Availability: OnlineOnly
Author: Antigravity

Extracts clean, readable content from web pages via the Defuddle CLI, cutting token usage versus WebFetch.

Get skill

Antigravity

Own this? Claim it

Spark score

out of 100

Updated 10 days ago

Version 15.7.0

Add to Favorites

Why it matters

This asset efficiently extracts clean, readable content from standard web pages, removing clutter like ads and navigation. It's ideal for tasks requiring token efficiency when processing articles, blog posts, or documentation.

Outcomes

What it gets done

Scrape and parse web page content.

Remove navigation, ads, and clutter from web pages.

Extract specific metadata like title, description, or domain.

Output content in Markdown or JSON format.

Install

Add it to your toolbox

Run in your project directory:

curl -fsSL https://spark.entire.vc/get/ag-defuddle | bash

Overview

Defuddle

Uses the Defuddle CLI to extract clean, readable content from web pages, removing navigation, ads, and clutter to reduce token usage compared to WebFetch. Use it when a user provides a normal webpage URL to read, summarize, or analyze, especially when token efficiency matters.

What it does

Uses the Defuddle CLI to extract clean, readable content from web pages, removing navigation, ads, and clutter to reduce token usage - preferred over WebFetch for standard web pages. Supports markdown output (--md, the default choice), saving directly to a file, JSON output containing both HTML and markdown, plain HTML with no flag, and extracting a single metadata property (title, description, or domain) via -p <name>.

defuddle parse <url> --md

When to use - and when NOT to

Use it when a user provides a normal webpage URL to read, summarize, or analyze - docs, articles, blog posts, and similar public web content - especially when token efficiency matters more than a noisier full-page fetch.

Inputs and outputs

Input: a public webpage URL, plus an optional output format flag (--md, --json, -o <file>, or -p <property>). Output: cleaned markdown by default, JSON with both HTML and markdown, raw HTML, or a single extracted metadata property. If not already installed, npm install -g defuddle sets it up.

Who it's for

Agents and developers reading or summarizing standard web pages who want clean content and lower token usage instead of a full noisy page fetch.

FAQ

Common questions

Discussion