Where your notes live decides
who really owns them.

Every notes app answers one question before it answers any other, usually without telling you: where does your writing actually live? The answer determines who can read your notes, what happens when the internet is down, what happens when the vendor changes prices, and how hard it will be to leave. Most people discover the answer only when something goes wrong. This lesson is about knowing it up front.

Two architectures, honestly compared

Broadly, notes live in one of two places.

Local files. Your notes are documents on your own disk, in folders you can see. Obsidian is the best-known example: a vault of plain Markdown files. The strengths are real: notes work with no account and no connection, any program can open them, backup is copying a folder, and no company can change the terms on which you access your own writing. The traditional weaknesses are also real: syncing across devices is your problem, and collaboration (two people editing at once, comments, permissions) does not come naturally to files.

A cloud database. Your notes are rows in a database run by the vendor, and the app is a window onto it. Notion is the best-known example. The strengths mirror the weaknesses of files: effortless sync, real-time collaboration, sharing with a link, access from any browser. The costs are the mirror image too: no connection often means no notes or a limited cache, the vendor's servers hold the only authoritative copy, and your access continues exactly as long as your account and their business do.

The lock-in problem

Lock-in is not a conspiracy; it is a gradient. Every hour you invest in a tool (pages written, structures built, links accumulated, teammates onboarded) raises the cost of leaving. That is fine and normal. It becomes a problem when the exit cost is not proportional to your investment but inflated by format: when the notes exist only in a proprietary shape that nothing else can read, so leaving means losing structure, links, history, or the notes themselves.

The test of a tool is not whether you will ever leave. It is what leaving would cost if you had to: the vendor is acquired, prices triple, the product pivots, or your organization mandates a change. Ten years is a long time in software, and notes are a decades-long asset. Your notes should outlive every app you ever write them in.

Open formats: what makes a format safe

An open format is one whose structure is documented and readable without the app that created it. Two matter most for notes.

Markdown is plain text with light conventions: # for headings, - for lists, **bold**. It has become the lingua franca of portable notes precisely because it is barely a format at all; every editor, every operating system, and every AI tool can read it. Its limitation is that it captures text structure, not rich structure: databases, per-block metadata, and layout do not fit naturally.

JSON carries structured data: typed blocks, properties, metadata. It is less pleasant to read raw, but it loses nothing, and any programming language parses it in one line.

NoteX Desktop uses both: notes are open JSON .notex files (metadata plus block content, readable without NoteX), and it edits plain .md files natively alongside them. Files sit in real folders on your disk, and everything exports to Markdown or self-contained HTML in bulk. The details are on files you own.

Local-first: a third way

For years the choice was binary: files and their friction, or clouds and their dependency. A newer school of software design, usually called local-first, refuses the binary. The principles are simple to state: the copy on your device is the primary copy, everything works offline, and the network is used for sync and collaboration when available rather than being required for basic function.

Done properly, local-first means the app is fully functional on a plane, the vendor disappearing does not take your notes with it, and sync improves the experience instead of gatekeeping it. NoteX Desktop is built this way: notes are files first, and there is an always-available Local workspace that never syncs at all, for material that should never leave the machine.

Sync as a choice, not a requirement

Sync itself deserves unbundling, because "sync" quietly bundles three separate questions: whether your notes replicate, where they replicate to, and who runs that destination.

Most cloud tools answer all three for you: yes, to our servers, run by us. A more honest design lets you answer each yourself. In NoteX, sync is opt-in per workspace, and the server is your decision: point the app at Lisan's cloud, or at any self-hosted Supabase-compatible server your team runs, keep several servers saved, and switch between them. A government team can keep notes on infrastructure it controls; a personal vault can stay purely local; a startup can use the hosted cloud and think about none of this. See workspaces and servers for how that works in practice.

How to evaluate any notes app for exit cost

You can run this audit on any tool, including NoteX, in about fifteen minutes. Do it before you invest years, not after.

• Export everything, today. Not one page: the whole workspace. If bulk export does not exist, that is your answer.
• Open the export without the app. Is it Markdown, HTML, or documented JSON that a text editor can read? Or a proprietary blob that only re-imports into itself?
• Check what survived. Page hierarchy? Internal links between pages? Images and attachments? Databases with their properties, or just flattened tables? Comments and version history usually do not survive anywhere; know that going in.
• Check the ongoing options. Is there an API that can read everything programmatically, so you can automate a continuous backup instead of relying on one heroic export at the end?
• Ask where the authoritative copy is. On your disk, on servers you chose, or only on the vendor's servers? What exactly do you have if your subscription lapses tomorrow?

The honest bottom line

If your top priority is real-time collaboration with a large team, a cloud architecture is genuinely better at it, and pretending otherwise would be marketing. If your priority is ownership, longevity, and independence, files are genuinely better, and local-first design now lets you keep most of the collaboration too. NoteX's position is to make ownership the default (files, open formats, your choice of server) and make sync and collaboration available on top, rather than the other way around.

Whatever tool you use: run the export audit, keep a periodic backup in an open format, and never let your only copy of a decade of thinking live exclusively on someone else's computer.

Next lesson: with the foundations settled, how to actually build something with them, starting with the team wiki.