Tana

Sweet spot: a long-form thinker who already lives in outliners (Roam, Workflowy, Logseq) and is willing to invest a weekend learning supertags. Once the schema clicks, Tana stops feeling like a notes app and starts feeling like a personal database that you happen to write into. Failure modes. The learning curve is real and it weeds out a lot of users in the first week — the empty Tana doc is genuinely intimidating compared to Apple Notes. Performance degrades on very large graphs, so heavy users start hitting cold-start and search lag at the 50k+ node range. The product's recent shift toward an AI-meetings pitch can feel disorienting if you signed up for the outliner. What to pilot. Pick one workflow you already do badly (weekly reviews, meeting notes, reading notes), recreate it in Tana with three or four supertags, and use it for two full weeks before deciding. If supertags are saving you re-typing the same fields over and over, you have found the win; if you are spending more time configuring tags than capturing notes, the tool is fighting your workflow and a simpler app is the right call.

Tana is an outliner-first notes app built around a single defining primitive: the supertag. Any node in your outline can be tagged (#meeting, #person, #task, #book), and that tag carries a schema — fields, default views, and behaviors. Tag a node #task and it gets a status field, a due date, and shows up in your tasks view automatically. The result is a knowledge graph that grows out of how you naturally take notes, not a database you have to design upfront. On top of supertags, Tana layers AI features: chat with your graph, voice-to-note transcription on mobile, AI commands that fill fields or summarise nodes, and a meetings module that captures audio, transcribes, and pulls action items into structured supertags. The product has been steadily shifting toward an AI-meeting workspace pitch through 2025–2026, but the underlying outliner-plus-supertag model is what long-time users stay for. Tana runs on Mac, Windows, web, iOS, and Android with a free tier (capped at 20k nodes, two workspaces, 500 AI credits) and paid Plus / Pro tiers at $10 / $18 per month. It sits in the same conceptual neighbourhood as Roam Research and Logseq, but the supertag schema layer makes it feel closer to a hand-coded Notion than to a pure outliner.