Extract clean Markdown from web pages with the Defuddle CLI — strips nav, ads, and clutter to save tokens. A token-efficient alternative to raw WebFetch for standard pages.
---
name: Defuddle
description: Extract clean Markdown from web pages with the Defuddle CLI, stripping navigation, ads, and clutter to save tokens. Use instead of raw WebFetch when given a URL to read or analyze (articles, docs, blog posts). Do NOT use for .md URLs — fetch those directly.
---
Use the Defuddle CLI to pull clean, readable content from standard web pages, removing boilerplate so you spend fewer tokens reading.
## How to use
1. Install if needed: `npm install -g defuddle`.
2. Get Markdown (the default choice): `defuddle parse <url> --md`.
3. Save to disk: `defuddle parse <url> --md -o content.md`.
4. Pull a single metadata field: `defuddle parse <url> -p title` (also `description`, `domain`).
5. Need both HTML and Markdown? Use `--json`; omit format flags for raw HTML.
Reach for Defuddle over WebFetch on normal HTML pages; for URLs that already end in `.md`, fetch them directly instead.
Full skill & source: https://github.com/kepano/obsidian-skills/tree/main/skills/defuddleSign in to rate and review this skill.
No reviews yet. Be the first to review this skill.