Most people who try the zettelkasten method quit inside a month. They read about Niklas Luhmann's 90,000-note slip box, set up a folder in Obsidian, write ten notes, and then stall on the part nobody warns them about: the linking, the tagging, and the weekly upkeep that the whole system depends on. The method works. The manual maintenance is what breaks.
This guide explains the zettelkasten method the way it actually functions, gives you a copyable note template you can use today, and shows where AI removes the parts that cause most people to abandon it.
Key takeaways
- The zettelkasten method is a network of small, self-contained notes ("atomic notes") connected by explicit links, not a folder hierarchy.
- Three note types drive it: fleeting notes (raw capture), literature notes (from what you read), and permanent notes (your own restated ideas).
- The value is in the links between notes, not the notes themselves. A note with no connections is a dead end.
- The system fails from maintenance load: manual IDs, manual linking, and manual review are a part-time job.
- AI speeds up the mechanical parts (suggesting links, surfacing related notes, extracting ideas) so you spend time thinking, not filing.
- Tools like Obsidian give you the vault; the open problem is keeping it current without a weekly review habit.
What Is the Zettelkasten Method
Zettelkasten is German for "slip box." The method comes from Niklas Luhmann, a German sociologist who wrote roughly 70 books and hundreds of articles using a paper card index of interlinked notes. Each idea got its own card, each card got a unique number, and cards pointed to other cards by number. Over decades this turned into a dense web he could query like a research partner.
The digital version keeps the principles and drops the index cards. The core rule is simple: one note holds one idea, and every note links to the notes it relates to. You are not building a filing cabinet. You are building a graph.
That distinction is the whole game. In a folder system, a note lives in exactly one place and you find it by remembering where you put it. In a zettelkasten, a note can be reached from any related idea, and the connections themselves become a source of insight. You discover links between things you wrote months apart, and those links are where new thinking comes from.
The Three Note Types (and Why They Matter)
The zettelkasten methodology runs on three kinds of notes. Skipping any of them is the most common reason a slip box turns into a junk drawer.
- Fleeting notes. Quick, throwaway capture. A thought on a walk, a line from a podcast, a reaction to a meeting. These are temporary. You process them within a day or two and then delete them.
- Literature notes. What you take from a specific source: a book, paper, or article. Written in your own words, with the source cited. One idea per note. If you copy a quote, mark it clearly as a quote.
- Permanent notes. The heart of the system. You restate an idea fully in your own words, as if explaining it to someone else, and connect it to existing notes. These are written to last and to be reused.
The flow is: capture as fleeting, read into literature notes, then convert the good material into permanent notes that link into your existing web. Fleeting and literature notes are raw input. Permanent notes are the asset.
Here is a real permanent note you could paste into any Markdown app today:
# 202607231430 Atomic notes force clearer thinking
An atomic note holds exactly one idea. The constraint is the point:
if a note needs two headings to explain, it is really two notes.
Splitting ideas forces you to name each one, which is where the
actual thinking happens. A vague note you cannot title is a vague
idea you have not finished having.
Links:
- [[202606190900 One idea per unit reduces coupling]]
- [[202607010815 Naming a thing is understanding a thing]]
Source: Ahrens, How to Take Smart Notes (2017)
Tags: #zettelkasten #note-taking #thinking
The number at the front is the unique ID (here, a timestamp: year, month, day, hour, minute). The [[double brackets]] are wikilinks to other notes. Tags help, but links do the heavy lifting.
How the Zettelkasten System Works Step by Step
You do not need special software to start. Any app that supports plain text and links will do. Here is the loop.
- Capture fleeting notes anywhere. Phone, paper, a voice memo. The goal is zero friction, not neatness.
- Read into literature notes. When you read something worth keeping, write one note per idea in your own words, and cite the source.
- Write permanent notes. Take the best literature and fleeting material and restate each idea as a standalone permanent note. This is where you slow down and think.
- Link deliberately. Before you save a permanent note, ask which existing notes it connects to. Add those links. A note with no links is a note you will never find again.
- Build entry points. As clusters form, create index or "structure" notes that link to the main threads of a topic, so you have a way in.
- Review and prune. Delete processed fleeting notes. Fix broken links. Merge duplicates.
Steps 4 and 6 are where nearly everyone falls off. Deciding which links to add means holding your whole vault in your head, and pruning is unpaid janitorial work. The bigger your slip box gets, the heavier both become. That is the maintenance tax, and it is the single reason most zettelkasten setups go quiet after a few weeks.
Zettelkasten in Obsidian: What Works and What Stalls
Most people run zettelkasten in Obsidian, and for good reason. It stores notes as plain Markdown files on your machine, supports [[wikilinks]] and backlinks natively, has full-text search, and renders a graph view so you can see your web of notes. For the zettelkasten obsidian pairing, the mechanics are a strong fit.
What Obsidian does not solve is the human upkeep. It shows you backlinks, but it does not decide which links a new note should have. It stores your notes, but it does not notice when two notes are really the same idea, or when an old note now contradicts a newer one. The vault is only as current and connected as your last manual pass through it.
| Task | Manual zettelkasten | Obsidian | Where AI helps |
|---|---|---|---|
| Store atomic notes | Index cards | Markdown files, local | Same files, read directly |
| Create links | Write IDs by hand | Type [[wikilinks]] | Suggest links from note content |
| Find related notes | Flip through cards | Backlinks + search | Surface related notes by meaning |
| Extract ideas from reading | Copy by hand | Copy by hand | Draft literature notes from a source |
| Spot duplicates or conflicts | Memory | Manual review | Flag overlapping or outdated notes |
| Keep the web current | Weekly review | Weekly review | Runs continuously in the background |
The pattern is clear. Obsidian removes the paper. AI removes the busywork on top of it.
How AI Speeds Up the Zettelkasten Method
AI does not replace the thinking. Writing a permanent note in your own words is the point, and no model should do that for you. What AI removes is the mechanical layer that surrounds the thinking: finding links, resurfacing forgotten notes, extracting first-draft ideas from a source, and noticing when something has gone stale.
Three concrete accelerations:
- Link suggestion. Instead of you recalling every related note, an AI reads your new note and surfaces the ones it connects to by meaning, not just by matching keywords. You approve the links; you do not hunt for them.
- Related-note recall. When you sit down to write about a topic, the system pulls the notes you already wrote that touch it, including ones you forgot existed. This is the "conversation with your slip box" that Luhmann described, without the flipping.
- Reading into notes. Point AI at a PDF or article and get a first draft of literature notes, one idea per note, which you then edit into permanent notes. You still do the restating, but from a running start.
The honest boundary: AI is good at the mechanical and the retrieval. It is not good at deciding what matters to you. Keep the judgment; hand off the filing.
There is a second, quieter problem AI can address. A zettelkasten is only useful if it stays current, and a plain vault does not update itself. When your view on something changes, the old note just sits there, wrong. Systems that track when a fact is superseded (marking the old version as outdated while keeping the history) solve the exact failure that kills long-running note systems. That "stays current" behavior is what a static folder of Markdown can never do on its own. If you want the deeper version of that argument, we wrote it up in how to give your AI a memory that actually lasts.
Fleeting, Literature, and Permanent: A Side-by-Side
If you remember nothing else, remember which note is which. Most abandoned zettelkastens are just a pile of fleeting notes that never got processed.
| Note type | Lifespan | Written in your words? | Linked? | Example |
|---|---|---|---|---|
| Fleeting | Days | Rough, optional | No | "podcast: spaced repetition works for concepts too?" |
| Literature | Kept with source | Yes, one idea each | Sometimes | "Ahrens argues writing is thinking, not recording it" |
| Permanent | Indefinite | Yes, fully | Always | "Writing forces you to notice gaps in an idea you thought you understood" |
The move that makes the whole method pay off is turning literature and fleeting notes into linked permanent notes on a regular basis. Do that, and the slip box compounds. Skip it, and you have a graveyard.
Where Dictation Fits
A lot of your best raw material never gets written down because typing it out is friction. Dictation closes that gap. Speaking a fleeting note into your notes while you walk, or narrating a rough permanent note and cleaning it up later, keeps the capture step fast enough that you actually do it. Tools like Contextli turn spoken thoughts into text notes that can feed straight into your slip box, so the capture friction that kills most zettelkasten habits mostly disappears.
Keep the Judgment, Automate the Filing
The zettelkasten method is one of the most durable thinking systems ever built, and the digital tools around it are excellent. The gap has always been maintenance: the linking, the reviewing, the noticing what has gone stale. That is exactly the layer worth handing to software.
Locul is a local-first desktop app that reads your notes where they already live (Obsidian, Notion, files, dictation) and builds a searchable, self-updating second brain on top of them, then serves it to your AI tools. It is free to start, 500 memories, local AI, so your slip box can stay connected and current without becoming a second job. You keep writing the permanent notes. It keeps the web wired together. If that sounds like the missing piece, you can download it here.
---
FAQ
What is the zettelkasten method in simple terms?
It is a note-taking system where you write small notes that each hold one idea and link them to each other, so your notes form a connected web instead of a folder tree. Over time the links between notes become a source of new ideas. It comes from sociologist Niklas Luhmann, who used a paper card index to write dozens of books.
What is the difference between fleeting, literature, and permanent notes?
Fleeting notes are quick temporary captures you process and delete within a day or two. Literature notes are ideas taken from a specific source, written in your own words with a citation. Permanent notes are fully restated ideas linked into your existing web, written to last. The permanent notes are the real asset; the other two are raw input.
Is zettelkasten better in Obsidian or on paper?
For most people, digital wins because linking, backlinks, and full-text search are instant, and your notes stay searchable as the collection grows. Obsidian is a common choice since it stores plain Markdown locally and supports wikilinks natively. Paper still works and forces slower thinking, but it does not scale to thousands of interlinked notes the way software does.
Can AI write my zettelkasten notes for me?
It should not write your permanent notes, because restating an idea in your own words is the thinking step that makes the method valuable. AI is best at the mechanical layer: suggesting links, surfacing related notes you forgot, and drafting literature notes from a source you then edit. Keep the judgment, automate the filing.
How is a zettelkasten different from Building a Second Brain?
Building a Second Brain (the PARA and CODE approach) organizes notes around actionable projects and areas, optimized for getting things done. Zettelkasten organizes notes around ideas and their connections, optimized for thinking and writing. They can coexist: use project folders for execution and a linked slip box for developing ideas.
Why do most zettelkasten systems get abandoned?
The maintenance load. Deciding which links a new note needs, pruning duplicates, and running a weekly review are real ongoing work, and it grows with the size of the collection. Most people stall on the linking and reviewing, the collection stops being current, and it quietly becomes a junk drawer. Offloading the mechanical upkeep is what keeps a slip box alive. For a related take on stopping frozen collections from rotting, see what a memory pack is.
FAQ