Skip to content
Primores AI wiki

Zettelkasten — What It Means

Zettelkasten

TL;DR: Zettelkasten is a note-taking method where each note contains one idea and links to related notes. Over time, connections emerge organically. It’s the ideal structure for LLM-maintained wikis because each note is atomic and self-contained.

Simple Explanation

“Zettelkasten” is German for “slip box” — a system invented by sociologist Niklas Luhmann who used it to write 70+ books and 400+ articles.

The core principles:

  1. One idea per note — Keep notes atomic
  2. Link notes together — Every note connects to related notes
  3. Let structure emerge — Don’t force categories; let connections create structure
  4. Write for your future self — Each note should be understandable standalone

Instead of organizing by folders/categories, you organize by connections.

Why It Matters for AI Knowledge Bases

Zettelkasten is the ideal structure for LLM-maintained wikis:

Zettelkasten PrincipleLLM Wiki Benefit
Atomic notesEach note can be independently updated
Links between notesLLM can auto-generate cross-references
Emergent structureWiki organizes itself as it grows
Self-contained notesEach note is citable by AI engines

Real-World Example

The Telegram Community Wiki Bot uses Zettelkasten:

[participant/alex.md]
# Alex K
Expertise: SEO automation, SaaS growth
Projects: projects/growth-bot
Mentioned in: discussions/2026-04-05-traffic


[projects/growth-bot.md]
# Growth Bot
Creator: participant/alex
Tools used: tools/perplexity, tools/python
Discussion: discussions/2026-04-05-traffic


[tools/perplexity.md]
# Perplexity
Used by: participant/alex, participant/maria
For: Competitor research, content ideas

Everything links to everything. Finding related knowledge = following links.

Zettelkasten vs Traditional Folders

Traditional FoldersZettelkasten
Note lives in ONE folderNote links to MANY notes
Structure is predeterminedStructure emerges
Finding = search or browseFinding = follow connections
Hard to maintainSelf-organizing
One-dimensionalMulti-dimensional

For LLM Implementation

When an LLM maintains a Zettelkasten wiki:

On ingest:

  • Create atomic notes for each distinct concept
  • Add links to existing related notes
  • Update existing notes with new links

On query:

  • Start from relevant note
  • Follow links to gather context
  • Synthesize answer from connected notes

On lint:

  • Find orphan notes (no links)
  • Suggest missing connections
  • Identify redundant notes

Tools That Support Zettelkasten

  • tools/obsidian — Designed for Zettelkasten with graph view
  • Roam Research — Bidirectional linking
  • Logseq — Open-source alternative
  • Plain markdown — Wikilinks like this work anywhere

Common Misconceptions

  • ❌ Myth: Zettelkasten needs special software

  • ✅ Reality: Just markdown files with wikilinks work fine

  • ❌ Myth: You need thousands of notes for it to work

  • ✅ Reality: Benefits start from ~20-30 connected notes

  • ❌ Myth: It’s only for academic research

  • ✅ Reality: Works for any domain where knowledge accumulates

Key Takeaways

  • One idea per note, heavily linked
  • Structure emerges from connections, not folders
  • Perfect fit for LLM-maintained knowledge bases
  • Each note is atomic and independently updatable
  • Tools like Obsidian visualize the connection graph

Sources