Wikilinks vs tags vs folders: how to actually organize your second brain

*May 13, 2026 · 10 min read*

The "how do I organize my notes" question is the single most-asked question in every PKM community on the internet. Folders versus tags is the old argument. Folders versus tags versus wikilinks is the modern one. The honest answer in 2026 is that you need all three, but for different jobs — and most people get the assignment wrong.

This article maps each organizing mechanism to the job it's actually good at, walks through the failure modes when you use the wrong one, and lands on the practical layout that works in MindWiki (and any other markdown-vault tool that supports the same primitives).

TL;DR

Folders are for things with a finish line — projects, active areas of work. Use them sparingly.
Tags are for cross-cutting attributes — status: in-progress, confidence: low. Use them as YAML frontmatter properties, not as a primary navigation system.
Wikilinks are for the connective tissue between ideas. Use them aggressively. They're the load-bearing mechanism.

Most note-takers organize too hard with folders, too softly with tags, and never quite trust wikilinks to do the work folders pretend to do. The result is a fragmented vault that needs frequent re-orgs. The fix is to lean into wikilinks, demote tags to per-page properties, and treat folders as inboxes/areas — not as the primary index.

Folders are filing cabinets, not knowledge

A folder forces you to pick one home for a note. That's fine for things with a clear single home — an active project, a customer file, a course you're taking. It's actively bad for ideas that touch multiple domains.

Where folders work well:

projects/ — one folder per active project.
active/ or current/ — work that has a finish line.
capture/ — a single inbox where new material lands.
sources/ — one page per book / paper / talk.

Where folders fail:

Concepts. The note "hyperbolic discounting" doesn't have a single home — it's relevant to behavioral economics, to your dietary writing, to a customer research finding, to a literature review. Filing it once means three of those contexts can't find it.
Reference material. A snippet you'll re-read in five different contexts shouldn't live in one of them.
Long-running themes. Anything you'll write multiple notes about over years doesn't belong in any one folder.

The mistake people make with folders is treating them as the primary index — "I'll find this note by browsing to it." That works for the first 50 notes. After 500, you can't remember which folder you put something in, and a graph-first system wins.

Tags in note apps are usually one-dimensional labels: #productivity, #research, #draft. They're appended to pages and used as filters.

That's a fine model for *attributes* — properties of the note that you might want to filter on — but a bad model for navigation. Here's why:

A tag has no direction. #bayes on note A doesn't tell you A is related to anything specific in B. It just tells you both notes have something to do with Bayes.
Tags drift. You start with #research, you accidentally add #researchnote, three weeks later you're using #literature-review. The tag namespace is a mess by the end of year one.
Tags don't show context. A wikilink lives in a sentence; the sentence tells you *why* the link exists. A tag is decorative.

The right use of tags in 2026 is as properties in YAML frontmatter:

yaml

---
title: Hyperbolic discounting
area: slips
type: concept
tags: [behavioral-economics, decision-making]
confidence: medium
created: 2026-03-12
---

Frontmatter properties let you query — "show me every page with confidence: low and area: slips" — without becoming a navigation primitive. MindWiki's properties + table views are built around this exact model.

Wikilinks are the connective tissue

A wikilink — [[Page Name]] — is the bidirectional, contextual, in-sentence link that turns a folder of files into a knowledge graph.

Why wikilinks beat both folders and tags for ideas-that-cross-domains:

Bidirectional. The page you link to automatically knows it was linked from somewhere else. You navigate forward and backward without doing extra work.
Contextual. The link lives in a sentence. The sentence carries why the link exists.
Self-maintaining. Rename a page, every wikilink follows it (in tools like MindWiki that update links on rename).
Visible in a graph. A knowledge graph view turns the wikilink network into a map you can pan and zoom.

The hard part is the habit, not the mechanism. Most note-takers under-link by a factor of three or four. The right reflex is: every time you mention something you've thought about before, link it. Even if the target page doesn't exist yet — a ghost link is fine; when three notes link to the same ghost, you've discovered a page worth writing.

The practical hierarchy

The mental model that works in practice:

Folders        →  filing cabinet for active work
Wikilinks      →  connective tissue between ideas
Properties     →  attributes you want to query on

Reading those three lines as a workflow:

Folders for projects with finish lines. projects/q3-rebrand/. Once the project ends, archive the folder.
Wikilinks for everything else. Mention a concept, link it. Mention a person, link them. Don't pre-build folders for the things you link to — let the linked pages live wherever's convenient.
Properties for status. status: draft, confidence: low, area: slips, type: source. Use a properties view to filter, not a folder browser.

Worked example — a vault that actually scales

A vault for someone doing personal research + writing + a few active projects:

capture/              ← raw inbox
slips/                ← atomic concept notes (your real knowledge base)
sources/              ← one page per book / paper / talk you've consumed
people/               ← one page per person you've written about
projects/q3-rebrand/  ← active project (will archive when done)
projects/article-x/   ← active project (will archive when done)
maps/                 ← hand-curated index pages
_assets/              ← attachments

Notice what's *not* here:

No productivity/, philosophy/, psychology/ folders. Those are themes, and themes belong as wikilinks/properties, not folders. A note about hyperbolic discounting lives in slips/. The fact that it's about behavioral economics is a tag in frontmatter. The fact that you cited it from [[Sönke Ahrens]] lives as a wikilink.

This separation is what lets the vault scale to 5,000 notes without you having to rebuild the structure every quarter.

Where AI fits

A graph-shaped vault is what makes AI useful at scale, because the AI uses the same structure you do for retrieval:

mindwiki_search finds pages by content (keyword + vector hybrid).
mindwiki_similar finds semantic neighbors of a page — adjacent thinking you didn't manually link.
mindwiki_graph returns the wikilink graph as JSON so the AI can walk it.
mindwiki_list_pages filters by frontmatter properties (your confidence, area, type fields).

A folder-heavy vault with a flat note structure breaks all four of those. Wikilinks + properties is the geometry the AI wants too.

Common organizational mistakes (and the fix)

Mistake 1: deep folder hierarchies

projects/2026/q3/marketing/blog-posts/article-x.md. You'll forget where you put things by week three.

Fix: maximum two levels of nesting. projects/article-x/ is enough. The wikilinks make the rest discoverable.

Mistake 2: too many tags

#productivity #pkm #notes #knowledge #ai #second-brain #thinking #organization #workflow. You'll never reuse them consistently.

Fix: pick fewer than 15 high-level tag families. Treat tags as properties, not navigation.

Mistake 3: under-linking

You write a note. You mention a person. You don't link them. Three months later you're searching their name and you find nothing because the mention was buried in a paragraph.

Fix: mention something → link it. Even with a ghost link. The cost is one second; the discovery payoff is high.

Mistake 4: re-orging quarterly

You move every note around every three months because "the structure isn't working." It isn't working because you're treating folders as your primary index. The structure shouldn't need to change every three months. The links should.

Fix: stop re-orging. Lean into wikilinks. The folder layout you started with can almost always stay if the linking habit is healthy.

Mistake 5: pre-building empty folders

You set up philosophy/, psychology/, economics/ before you have any notes in them. Then you spend the next year wondering whether each new note belongs in two of them.

Fix: create folders only when you have material that demands them. Let structure emerge.

How MindWiki helps the habits stick

Wikilinks are first-class in the editor with live preview and rename refactoring.
Backlinks panel on every page surfaces what links into it.
Knowledge graph view (macOS app) shows the network so you can see clusters forming.
Frontmatter properties with table views on the macOS app turn properties into a queryable database.
Auto-Linker (Pro automation) proposes wikilinks where another page's title appears unlinked.
Pattern Detection (Pro automation) surfaces emergent clusters so you know when a topic has earned a folder.

The whole point is to make the right habits cheaper than the wrong ones.