Why people use tools like Tana and Roam
The main reason people jump into tools like Tana and Roam is because regular note-taking just completely crumbles once your thoughts start overlapping. Like, when your idea from one meeting needs to connect to a completely separate research summary you made three weeks ago while doomscrolling Reddit. Traditional folders? Useless. Tags? Start duplicating themselves like gremlins after midnight.
Roam was one of the first mainstream tools to convince people that maybe your notes don’t need to live in folders—or even inside documents, really. They could just live as blocks of thought, connected by backlinks and unstructured links. Like a wiki that auto-wires itself.
Tana comes into this space a little later, bragging about structure. Where Roam is like that messy-but-brilliant professor who only writes on whiteboards, Tana feels more like someone trying to turn your pile of post-its into a spreadsheet without telling you. It’s still super freeform—but under the hood, there’s a database hiding, even if you don’t notice it right away.
For me, I started out with Roam, used it for a year and a half, broke several setups, rage-quit, then slowly crawled back into documentation land using Tana—mostly because I wanted tagging that didn’t become tag soup 🙃.
Roam daily notes feel like a trap after a while
If you haven’t used Roam, the entire app basically tries to funnel you into one page every day: the Daily Notes page. At first it feels brilliant—just dump what’s in your head and move on, and connect stuff later. You’re like, wow, this is frictionless.
Until you start trying to find something you wrote three weeks ago… and realize you used three similar but inconsistent tags like #meetingNote, #Meetingnotes, and [[meeting-notes]]. And now you have no idea where anything lives.
Also, block references in Roam seem like a good idea—until you try to restructure content later. I once copied a set of notes for a client workshop into a different page to prep for version two. The original page? Still there. But every time I changed anything in the new version, it updated the old notes. Because what I’d copied was just more references to the original blocks.
Internally I screamed. There’s literally no warning. Copy-paste feels like it should detach things by default, but no. It’s like Roam just expects you to be emotionally ready for your prior work to morph behind your back ¯\_(ツ)_/¯.
Tana makes structure optional but nudges you anyway
The biggest difference I feel with Tana: it doesn’t force you to think in structure upfront, but it sort of gently nudges you if you’re that kind of person. It’s got a thing called “supertags” that feel like little templates you can assign to a note. So if I have an idea that’s actually a meeting, I can give it a supertag called “Meeting.” Now it has fields like date, participants, outcomes, and so on.
But here’s the wild part—you can build those yourself. It’s like if you started tagging your blog posts with #recipe and then suddenly got a prompt to include ingredients, cook time, and nutrition facts. Without breaking anything.
The first time I used Tana, I created a tag called #ScriptIdea just to keep track of dumb comedy ones I write for videos. After like four notes, I realized every one had the same pattern: premise, main gag, tone. So I turned #ScriptIdea into a supertag with those fields. And now every time I write one, Tana gives me those slots automatically.
In Roam, I would’ve been copying a checklist every time and praying I didn’t forget one.
This isn’t to say Tana is perfect though. Sometimes I’ll go to search my vault and realize it auto-completed to literally every item that has a similar name, even if that item is ten folders deep inside a random meeting note. It’s smart…but often too smart. I miss things because there’s almost too much context being pulled in.
The graph features feel different under the hood
Roam sells you hard on the idea that you’re building a vast network of thoughts. You can even click to see your idea map visually—a little dot map with nodes and connections. It gets messy fast, but it feels kinda magical. One note I made about stoicism started linking to three unrelated project plans because I used the same word in their summaries. 🤷♂️
In practice, though, Roam’s graph view is a bit of a gimmick. You can’t manipulate much in it. It’s more of a screensaver than a true navigable map.
Tana, by contrast, doesn’t show you a literal graph view right away. But every note is functionally a node in a tree, and you can nest nodes forever. Weirdly, this means that your whole vault is one big outline—and you can collapse and expand to reorganize your thinking.
One time I accidentally dragged a note about video editing tips into a nested node under my grocery list. For a week I thought I deleted it. Only while expanding the food planning branch did I find it. Tana’s infinite nesting is helpful—but it can absolutely let your content disappear by accident.
Search and link creation works totally differently
If you’re coming from Notion or Obsidian, both Roam and Tana will feel bizarre in how they treat links—especially auto-linking.
In Roam, putting [[double brackets]] around anything makes a page automatically. That page might be empty, but it now exists, with its own backlinks section. Good for quickly cross-linking—but also leads to too many ghost pages if you’re not careful. I once had a page called [[nap]] just because I wrote “maybe I need a nap” during a meeting.
Tana tries to avoid this clutter, but it means linking requires a bit more intent. You can tag a node or link by typing @nameOfNode, but it doesn’t create a page unless the node actually exists. And you can’t just free-form link everything with wild abandon.
This frustrated me at first. But then I realized I had far less junk content. In Roam, my sidebar was littered with half-linked fragments. In Tana, I could actually keep ideas centralized without accidentally spawning link hydras.
That being said, I *do* miss Roam’s real-time backlink inclusion. In Tana, links feel a bit more like cross-references inside a spreadsheet — they don’t give the same fuzzy magic of everything being indirectly connected.
Data export and backup give me anxiety in both
Okay, real talk: neither of these tools gives me the confidence of Markdown files stored in folders like Obsidian. Every time I tinker with workflows in Roam or Tana, a sliver of my brain screams “but where’s the raw data.” 😬
Roam does allow export to JSON and a version of Markdown. The issue is: it’s formatted weird. Like, it technically includes the structure, but it’s not clean Markdown you could just browse later. And block references export as separate items, with hard-to-parse UUIDs.
Tana exports let you download your whole workspace, but again, it’s structured—meant to be re-imported into Tana or something that understands the fields. If you’re just trying to get a plain-text version, you’re gonna cry. I tried pulling out one meeting summary and ended up with eight JSON fragments and no clear outline.
It’s clear that both tools don’t really expect you to leave. In fact, when I asked a friend to collab and they wanted to move our Roam notes into Notion, I literally gave up halfway. We just started over.
Real workflows I broke between both tools
Okay, this part is personal. Here are actual workflows I set up in these apps that broke, abused me emotionally, or required duct tape to fix:
– In Roam, I once created a “template generator” using buttons and queries. I reused it for three weeks… until I realized the query embedded static dates. Which means every new note was pulling in last week’s info.
– In Tana, I tagged several online tools I was testing with a supertag called #ReviewQueue. Everything worked… until I renamed the tag #AppReview. Now all the nodes still thought they were part of the old one. I had to rebuild filters and views manually 😑
– I tried to export a weekly journal from Roam to a Google Doc using a DOM-scraping script. Roam updated their layout and broke my entire automation. Zap dead. RIP Wednesday wrap-ups.
– In Tana, I built a “project dashboard” using live queries. But every time I marked something as complete, the query didn’t auto-refresh. So the finished projects hung around until I manually toggled the query off and back on. I thought it was broken. Nope. Just lazy refresh logic.
Which one plays nicer with other tools
Short version: neither is amazing here, but Tana is trying harder.
Roam has some decent community scripts floating around—but no official API that’s fully documented. You can use tools like Readwise to pull in highlights, but exporting structured content remains tough.
Tana is newer but started opening up its API and includes native importers for tools like Notion, Obsidian, and Roam itself. I tried importing a JSON export from Roam into Tana, and it kind of worked… except for styled blocks, embedded content, and dates. So yeah, still needs love.
The good part: Tana supports real custom views based on data filters, so you *can* make it feel more like Airtable or Notion if you’re willing to tweak. I’ve done setups where I create one view for tracking experiments, another that filters to only personal logs, and a third just to show which ideas I haven’t looked at in a week.
Roam, not so much. There’s no concept of views. Everything is just a long list of blocks and maybe a few queries. I love the flexibility—but hate the mess that follows.
Accidental complexity is easier to create than control
This is the thing I wish someone warned me about upfront.
Both Tana and Roam feel super lightweight at first. You open a blank page, type some thoughts, use a couple brackets, and you’re flying.
But give it three weeks. Suddenly every note is linked to five others. You’ve invented three types of tags without realizing it. Now you wish your notes had structure, but you’re four layers in.
In Tana, this shows up as too many custom fields you’re not even using. Like, I had a supertag for “Books to Read” that included Start Date, End Date, Author, Genre, Quick Thoughts, Highlight Type… and I filled out maybe two of those on any entry. Now it just looks messy.
In Roam, it’s over-linking. Pages multiply fast, and unless you’re constantly gardening the graph (yes, people call it that), it turns to soup. Search stops helping because everything looks “relevant.”
Flexible tools are only powerful if you design your own constraints. Otherwise, welcome to spaghetti world.
Tana tries to help some with templates and filtering views, while Roam just shrugs and offers you backlinks.