DEV Community

Qonspekt
Qonspekt

Posted on

Web clippers vs. Qonspekt: what's the difference for Obsidian users?

If you use Obsidian and want to save content from the web, you've probably tried a web clipper. Web2MD, WebInk, MarkDownload — they all do roughly the same thing: convert a webpage's HTML into Markdown.

That's useful. But it's not the same as taking notes.

What web clippers actually give you

You paste in an article URL, and you get back one Markdown file. The file is the article — its structure, its headings, its paragraphs — just in Markdown format instead of HTML.

What you don't get:

  • Identified concepts worth keeping
  • Atomic note structure
  • YAML frontmatter
  • Connections between concepts ([[wikilinks]])
  • Notes you can actually search and reuse

You still have to do the intellectual work: read the file, extract what matters, decide how to structure it, write atomic notes one by one.

What Qonspekt does instead

Qonspekt doesn't convert HTML. It reads the article and identifies what you'd actually want to keep in a Zettelkasten.

Input: a 3000-word article about the Feynman Technique.

Output:

  • feynman-technique.md — what it is, why it works
  • learning-by-teaching.md — the psychological mechanism
  • knowledge-gaps.md — how to find them
  • active-recall-vs-passive-reading.md — the contrast Feynman makes

Each note has YAML frontmatter, 2–4 tags, wikilinks connecting the concepts within the batch, and a ## Sources section. Download as ZIP, drag into Obsidian.

The key difference in one sentence

Web clippers preserve the article. Qonspekt writes notes about the article.

When to use which

Use a web clipper when:

  • You want to save a reference document for later
  • You need the full article structure preserved
  • You're archiving, not note-taking

Use Qonspekt when:

  • You want to add knowledge to your Zettelkasten
  • You're building an Obsidian vault of concepts
  • You read to learn, not just to collect

Practical comparison

Web Clipper Qonspekt
Output 1 Markdown file 3–7 atomic notes
Frontmatter No Yes (title, tags, aliases)
Wikilinks No Yes (between extracted concepts)
What it understands HTML structure Article concepts
Install required Browser extension No (browser-based, no install)
Cost per article Free ~$0.003 (Claude API, BYOK)
Backend None / server None (browser-only)

Try Qonspekt

qonspekt.github.io/qonspekt

Bring your own Claude API key. No account, no backend. New Anthropic accounts get free credits.

If you've tried other approaches to this problem — Readwise, Notion web clipper, custom scripts — I'd be curious what actually stuck for you.

Top comments (0)