The argument for plain text is usually made badly, as a matter of taste, by people who like terminals. The actual argument is duller and much stronger: plain text is the only note format with a track record.
A .txt file written in 1985 opens today, on hardware that did not exist then, in software written by people who never heard of it. Nothing else in personal computing has that property. Preservation institutions land in the same place for the same reason: when the Library of Congress ranks formats by how well they survive, unstructured text sits at the top, and proprietary application formats do not.
Your notes are a thirty-year asset if they are anything. That is the timescale the format question should be judged on.
What "plain text" actually means
Three properties, and only three:
- The file is readable without the app that wrote it. Open it in Notepad, in a terminal, in an email. The words are there.
- The formatting is visible characters, not hidden metadata.
**bold**rather than a binary flag pointing into a style table. - The files sit in a folder you can see. Not in a database, not in an app container, not exclusively on someone's server.
Markdown is the usual syntax, and CommonMark is the closest thing it has to a specification. The syntax is not the point though. A folder of .txt files with no formatting at all satisfies every property above.
What you actually get
Independence from the vendor. This is the whole thing. Every notes company either dies, gets acquired, or changes its pricing in a direction you dislike. When your notes are files, that is an inconvenience: you close one app and open another over the same folder. When your notes are in a proprietary store, it is a migration project, and migrations lose data.
Backup becomes trivial. Any tool that copies files backs up your notes correctly, including ones written before you were born. Version history is a git init away, which gives you a diff of what you believed last March. See how to back up your notes.
Search that cannot be taken away. grep and ripgrep search a hundred thousand notes in under a second and do not require a subscription, an index, or a network.
Automation. Scripts can read, write, rename, and transform your notes because there is nothing to reverse-engineer. Ten thousand notes become a dataset rather than a UI.
Real export. Not the export button that produces a zip of HTML you will never open. The files are the export, permanently, with no step in between.
What you give up, honestly
Plain-text advocacy usually stops before this section. It should not.
- Rich content is second class. Images become links to files. Tables are painful. Handwriting, audio, and PDF annotation are not text and never will be. If your notes are mostly sketchnoting or annotated PDFs, this entire argument does not apply to you.
- Sync is your problem. iCloud, Dropbox, Syncthing, and Git all work, and all can produce conflicted copies on a bad day. Managed sync in a closed app is genuinely more reliable, and pretending otherwise is dishonest.
- Mobile is worse. Almost universally. File-based apps on phones are clumsier than their desktop versions and clumsier than native mobile apps, and quick capture is where you feel it.
- Collaboration is weak. Two people editing the same Markdown file is a merge conflict waiting to happen. For shared work, use a tool built for it. See the best notes app for teams.
- No structured data. Databases, relations, rollups, and views are the reason people use Notion, and they do not exist in a folder of text.
- The tinkering trap. Plain text attracts people who enjoy configuring things, and configuring a system feels like using it. Watch for this. It is the most common way a plain-text setup produces no notes at all, and it is a close cousin of the collector's fallacy.
The rule that makes the trade-off manageable
You do not have to choose once and forever. The useful principle is narrower:
Keep the notes you want in ten years as files. Keep everything else wherever it is convenient.
Most people's notes split cleanly. There is a small permanent core, the ideas and reference material and writing that would genuinely hurt to lose, and a much larger body of grocery lists, meeting scraps, and half-thoughts with a shelf life measured in weeks. The core deserves a durable format. The rest does not, and forcing it into one is how people end up with an elaborate system for remembering to buy milk.
In practice: capture wherever is fastest, promote what survives into text files. The promotion step is the same one described in the note inbox, and it is the moment when fleeting notes become permanent notes.
Making it actually durable
A folder of Markdown is necessary but not sufficient. Four habits do the rest:
Write the metadata in the file. Tags, dates, and source in the text itself, not in a sidebar the app owns. Anything stored outside the file does not survive the app.
Prefer app-neutral syntax. Wiki links, callout blocks, and query syntax are app features wearing Markdown's clothes. They are fine, and you should know which parts of your notes stop making sense elsewhere.
Name files so they still make sense stripped of everything. A folder of Untitled 47.md is technically plain text and practically worthless. See how to name notes.
Test the claim once a year. Open your notes folder in a plain text editor with the app closed. If it is unusable, the promise was theoretical. This takes five minutes and is the only real audit available.
Who should ignore all of this
If you are on iPhone and iPad only, capture on the move, and want notes to be effortless rather than owned, the file-based route will cost you more than it returns. Use something native and keep a real export habit instead.
The format question only starts to pay once your archive is large, old, and load-bearing. Below that threshold, the app that you will actually open beats the format that will outlive it.
More in the best markdown notes app, digital versus paper notes, the best offline notes app, and the wider personal knowledge management section.