Note-Taking Apps for Thinking, Not Hoarding

Everybody evaluates note apps on capture speed. Capture is the easy part — every app on the market gets a thought out of your head in under three seconds.

The interesting question is what happens six months later, when you half-remember writing something about a supplier and you need it now. Could your current setup find it?

Retrieval is the actual product

Three retrieval strategies exist, and they suit different people:

• Search. Works if you can remember a distinctive word. Requires nothing from you at write time. Fails on notes about common topics.
• Links. Works if you connected the note to something you will visit again. Requires a small habit at write time. Surfaces things you were not looking for, which is either the main benefit or a distraction.
• Structure. Folders and tags. Works if your categories were right at the time. They usually were not.

In practice: search handles most retrieval, links handle the valuable surprises, and structure handles almost nothing. Most people invest their effort in exactly the reverse order.

A tagging scheme is a prediction about what future-you will be looking for. That prediction is usually wrong, and maintaining it costs real time every week.

The portability question

Note apps are a category with unusually high switching costs and unusually short company lifespans. A vault of five years of notes in a proprietary format is a liability.

The test is simple: can you take your notes and leave, today, without a converter? Markdown files in a normal folder pass. Anything requiring an export button that produces a zip of JSON does not, in practice — technically portable, practically abandoned.

Where local-first apps struggle

Honesty about the trade-off: local-first apps are meaningfully worse at sync, collaboration and mobile. If you write from a phone as often as a laptop, this is not a minor detail. Cloud-first apps are better at this and always will be.

The reasonable middle ground most people land on: local markdown files, synced by a general-purpose file sync service. Slightly clunky, no lock-in.

What to actually check in a trial

1. Type a note, close the app, reopen it. How many seconds to a blank note?
2. Search for a word inside a note you wrote a week ago. Instant, or a spinner?
3. Open the storage folder. Can you read the files without the app?
4. Write on your phone. Is it tolerable, or a worse version of the desktop app?

Four checks, ten minutes, and they predict a year of use better than any feature comparison.

Why the elaborate methodologies keep failing

Every few years a note-taking methodology becomes popular, acquires an acronym, and generates a large amount of content about how to set it up. The setup content vastly outnumbers accounts of anyone still using it two years later, and that ratio is the finding.

The failure is structural rather than a lack of discipline. These systems front-load work at capture time — deciding a category, adding links, filing into a hierarchy — in exchange for a benefit at retrieval time that arrives months later, if at all. Under pressure the cost is immediate and the benefit is theoretical, so the cost is what gets cut. What remains is a half-maintained system, which retrieves worse than no system at all, because you now trust it and it is incomplete.

The systems that survive invert this: near-zero cost at capture, and an acceptance that retrieval will mostly be search.

Notes that turn out to matter get one extra step

The exception to minimal capture is small and worth making. Roughly one note in twenty turns out to be load-bearing — a decision and its reasoning, a supplier's terms, why something is built the way it is.

Those deserve two minutes: a descriptive title containing the words you would actually search for, and a link from wherever you will be when you next need it. Not a tag. A link from the project note, the person, the recurring document — the place your future self will already be standing.

You cannot identify these at capture time, which is exactly why the weekly ten-minute pass exists. It is not filing. It is asking, of the last week of notes, which two will still matter in six months.

A note on AI features in note apps

Every note app has added AI search and summarisation. The summarisation is largely irrelevant — your own notes are already short, and a summary of a summary loses the specific detail that was the reason to write it down.

Semantic search is genuinely useful, because it addresses the exact failure of keyword search: finding a note when you remember the concept but not the words. Treat it as a better search box rather than a reason to switch tools, and check how it works if your notes are private — some implementations send note content to a third-party service, which for a personal journal or client notes is a decision worth making deliberately rather than by accepting a default.

The system that survives a busy week

We are deliberately not naming a winner here, because retrieval style is personal and the habit matters more than the app. Every elaborate note-taking methodology works in a calm week. The test is a week where you are behind on everything. What tends to survive is: one inbox, dated notes, links when they are obvious, and a weekly ten-minute pass to file anything that turned out to matter. Anything more elaborate than that gets abandoned, and an abandoned system is worse than a simple one. The same test applies to the task manager next to it, for the same reason.