Tana Review: The Logic of Relational Context and the Intellectual Sovereignty Unhack

You had the idea. You’re sure of it. Some Tuesday eighteen months ago you wrote down something genuinely sharp — a connection, a phrasing, a way through a problem you’re now facing again. You can almost feel the shape of it. But it’s somewhere in the 1,400 notes you’ve accumulated across three apps and a graveyard of folders, and you will not find it. So you sit there and reinvent a worse version of your own thought, again, paying twice for an insight you already had.

The short version: Tana is a note-taking app built on a graph database instead of folders. Its central idea is the Supertag: when you tag a note as #person or #project, you don’t just label it — you give it a structure (a #person expects an email and a role; a #project expects a deadline and an owner) and connect it to everything related. That lets you ask real questions of your notes — “show me every task tied to this project due next week” — that a folder system simply can’t answer. It suits people doing heavy research, writing, or strategic thinking who need to retrieve ideas months or years later. It costs roughly $10–15 a month, takes two to four weeks to get fluent in, and is weak at team collaboration and true offline use. If you take casual notes you rarely revisit, it’s overkill.

What is Tana, and how is it different from a normal note app?

Tana is a knowledge tool where every piece of information is a “node” in a connected graph rather than a file in a folder. Most apps — Evernote, Apple Notes — work like a filing cabinet: you make folders, drop things in, and hope you remember where. It works fine until it doesn’t, and the failure point is always the same. You end up with a thousand notes and no way to surface the handful that matter right now.

Tana flips the question you’re answering. In a folder app, every new note forces you to decide where do I file this? In Tana you instead answer what is this? — you tag it #task or #meeting or #book — and the system handles the organisation from the tag. A note tagged #person already knows it should hold an email, a role, a last-contact date. The structure isn’t something you maintain on top of your notes; it emerges from how you label them.

The shift is small to describe and large to live with: you stop sorting information and start describing it, and the connections build themselves.

What problem does Tana actually solve? Information amnesia

Here’s the pain underneath the whole category, and it has a name worth giving it: information amnesia. You don’t lose notes because you’re disorganised. You lose them because folders are silos by design — each note sits in one place, disconnected from every related thought, its context quietly decaying until the day you need it and can’t reconstruct the path back.

The reframe that makes Tana click is this: the folder was never neutral storage — it was actively severing the connections that make your notes worth keeping. Every time you filed an idea under one heading, you cut it off from the five other ideas it belonged with. You weren’t building a second brain. You were building a thousand sealed rooms and losing the keys one by one.

A graph removes the walls. Because any node can link to any other, your notes stop being a pile and become a network you can interrogate. Ask “which books I read this year connect to this theme” and the answer assembles itself from links you made in passing. Your thinking becomes cumulative — each note compounds on the last — instead of a treadmill of rediscovering what you already knew.

What is a Supertag, and why is it the core of Tana?

The Supertag is the feature everything else hangs on, so it’s worth getting concrete. A Supertag is a tag that carries a structure. Create a #project Supertag and you define its required fields once — deadline, budget, owner, status. From then on, every node you tag #project sprouts those exact fields, automatically. No more half-finished entries where one project has a deadline and the next doesn’t.

The payoff is that complexity stops breaking your system. In a folder world, more stuff means more chaos and no folder hierarchy digs you out. With Supertags, structure is inherited: add a new field to #project once and it propagates to every existing project at the same time. And because every tagged node is structured data, you can run queries that would be impossible in any filing cabinet — “every #task assigned to this person, tied to this project, due in the next seven days” returns an actual list, not a folder you have to read through.

A Supertag turns a note from a scrap of text into a small database record — and a pile of records is something you can finally ask questions of.

How fast is Tana, and can you get your data out?

Tana keeps your whole graph cached locally in the browser, so search and navigation run on your machine rather than waiting on a server. The practical result is that retrieval stays near-instant even as you pass tens of thousands of nodes — comfortably into the 50,000–100,000 range on a normal modern computer. Most note apps get slower the more you put in them; this is the rarer case where scale doesn’t punish you. The honest caveat: it leans on your hardware, so a very old machine with little memory will lag. For most people that’s a non-issue.

On the lock-in question that should worry anyone trusting an app with years of thinking — Tana exports your graph to detailed JSON, so your data isn’t hostage. That said, be clear-eyed: an export is not a clean migration. The relational structure you build doesn’t drop neatly into another tool, so the more you customise your schema, the more real the switching cost becomes. You can leave with your data; you can’t leave with your system intact.

There’s also a privacy angle, stated carefully. Tana processes locally first and its AI features operate within the structure you’ve defined rather than treating your notes as raw text to mine. That’s a genuinely better posture than the industry default — though “better than default” is the honest claim here, not “zero-knowledge fortress.” Treat any cloud note app as cloud, and keep your most sensitive material somewhere built for secrecy.

How to set up Tana so it stays clear: a starter system

The relief is that you don’t architect a cathedral on day one. You start with five tags and let the rest earn its place.

Begin with the “Big 5.” Tana’s commonly recommended starter ontology is just five Supertags: #task (a unit of work with a deadline and owner), #project (related tasks with a timeline), #person (contacts with roles and history), #resource (books, articles, tools you want to find again), and #note (loose thoughts that may later grow into one of the others). Give each one at least a single mandatory field from the start — that one constraint forces consistency before bad habits set in. You’ll refine these after a month of real use; nearly everyone does.

Run a daily node. Make one node dated for today and hang the day’s captures under it. Over a month those daily nodes become a searchable record of your own thinking.

Capture without friction, process later. Tana’s Capture is a quick mobile and voice inbox: a fleeting thought goes in raw in seconds, and you tag and connect it later. Ideas die in the gap between having them and recording them, and separating capture from processing closes that gap.

Audit monthly. Use a live query to find untagged or half-filled nodes and either fix or bin them. A graph stays useful only if you don’t let dross pile up.

How Tana’s capture and AI features fit the workflow

Two features turn the graph from a tidy database into something you’ll actually use under pressure. The first is Capture, Tana’s quick-input layer for phone and voice. The point of it is friction: an idea has a half-life, and the seconds between having a thought and recording it are where the detail leaks out. Capture lets you dump the raw thought — typed or spoken — into an inbox in a couple of seconds, no navigating, no deciding where it goes. You tag and connect it later, when you’re at a desk and thinking clearly. That deliberate split between capturing and processing is what lets you think fast in the moment without the nagging job of filing.

The second is the AI layer, which can run a model directly against your nodes — summarise a project, turn a meeting note into a task list, pull the through-line out of a cluster of related notes. The meaningful difference from pasting text into a general chatbot is that Tana’s AI works inside the structure you built: it knows what a #project is in your system and which fields matter, so it respects your ontology instead of guessing at flat text. It’s the difference between an assistant who has read your filing system and one you have to brief from scratch every time. Useful — and, per the privacy note above, something to use with the same judgement you’d apply to any cloud AI feature.

Frequently asked questions

Is Tana better than Obsidian?
Neither is “better” — they’re built for different jobs, and pretending otherwise leads you to the wrong tool. Obsidian is a local-first Markdown editor whose strengths are total data ownership and easy portability; your notes are plain files you fully control. Tana is a relational graph database whose strengths are structured queries and inheritance. Choose Obsidian for simplicity and ownership, Tana for semantic relationships and asking questions of your notes. See our Obsidian review if portability is your priority.

Can I use Tana for team projects?
You can, but it’s clunky, and you’ll feel it. Tana’s sharing exists but isn’t as polished as a tool built for collaboration. For real-time shared workspaces and project management, Notion is the better fit; Tana shines for solo knowledge work and only tolerates the occasional shared query.

How much data can Tana handle?
Comfortably 50,000–100,000 nodes on a modern machine before you’d notice any lag, and the real ceiling is higher. Practical performance tracks your hardware more than the node count. Most solo users never hit a wall.

What happens if I change my tag structure after capturing thousands of nodes?
You can restructure, but it’s manual — there’s no painless global remap. The practical advice is to plan your core handful of Supertags carefully in the first week or two, then refine gradually, rather than expecting to overhaul the whole ontology six months in.

Does Tana work offline?
Partially. It caches your graph locally so you can view and search without a connection, and captures sync when you reconnect — but it isn’t truly offline-first the way Obsidian’s plain files are. It’s fine for patchy connectivity, not for working off-grid for days.

You opened this still half-believing that lost Tuesday idea was gone for good, and resigned to reinventing your own thoughts on a loop. That loop was never a memory failure on your part — it was the predictable cost of storing connected ideas in disconnected folders. Tana’s bet is simple: describe what each thing is, let the links form, and your notes turn from a junk drawer into something you can actually question. It isn’t for everyone, and this review has named exactly where it isn’t — casual notes, team work, true offline. But if you live at the meeting point of high information volume and thinking-in-connections, the right tool stops being an archive you dig through and becomes a partner that hands ideas back to you before you’ve finished asking. You’re not someone who keeps losing their best thoughts. You’re someone about to stop.

Related reading: Building a Second Brain Review, MasterClass Review, NextDNS Review, and Superhuman Review.