Logseq vs Tana: which outliner in 2026?
Tana vs Logseq is the outliner-versus-outliner choice: both start from daily notes and nested bullets with references, then part ways on everything else. Tana Outliner is a cloud workspace whose supertags turn bullets into typed, queryable objects, with AI metered by credits on every plan; Logseq is free, open source and local, and its 2.0 database version — beta since July 2026 — added the typed properties and Tana-style tags that used to be the reason to pay for Tana. Every price and version below is dated.
The short answer: Logseq if free, open source and a graph you own — files, or a local database — matter more than polish, and you can ride the 2.0 beta or park on the classic app. Tana if supertags with live queries, the meeting and voice agents and a hosted multiplayer workspace are worth $8–14 a month for Tana Outliner. Logseq DB has narrowed the structure gap; it has not closed the AI, hosting or finish gaps — and those are exactly what Tana charges for.
- Choose Logseq if you think in daily notes and block references, want an AGPL-licensed app with your graph on your own device, and your structure needs are tags with a few properties rather than a schema.
- Choose Tana if supertags, fields and live queries fit how you think, you’ll use the meeting transcription and voice capture, or you need a real-time shared workspace — and metered AI credits are acceptable.
| Logseq | Tana | |
|---|---|---|
| Price | Free; official sync via project backing during beta | Plus $8/mo, Pro $14/mo (22% off yearly); students, academics, NGOs 50% off |
| Free plan | The whole app, no limits | Yes — 500 AI credits/mo, 3 workspaces, 0.5 GB |
| Where notes live | Markdown/Org files (classic) or a local database (2.0) | Tana’s cloud; no self-hosting, no end-to-end encryption |
| Structure | Tags with typed properties (2.0); block refs, queries | Supertags: typed nodes with fields, inheritance, live queries |
| Built-in AI | None — bring your own via plugins | Meeting agent, voice agent, AI chat — credit-metered |
| AI agents / MCP | Community MCP servers; no official one | Local MCP server on by default (free); hosted server in beta |
| Offline | Always — a local app | Desktop app; personal workspaces editable offline since late 2025 |
| Export | The files are the export (classic); Markdown export (2.0) | Workspace to Markdown or JSON; supertags and fields flatten to text |
As of November 2026 — Logseq 2.0.2 DB beta (Oct 7, 2026; nightly Oct 25; classic 0.10.15, Dec 2025) per GitHub releases; Tana Outliner plans, local MCP server and export per outliner.tana.inc, re-read Nov 7, 2026.
What is Logseq?
Logseq is the open-source counterpart to Roam: an outliner built on bidirectional links and daily notes, but storing everything as local Markdown (or Org-mode) files on your device. It’s free, AGPL-licensed, and adds flashcards with spaced repetition, PDF annotation, queries, whiteboards and a plugin ecosystem. Sync is available end-to-end encrypted, or you can roll your own with Git or Syncthing.
The project is mid-transition: alongside the classic Markdown version, a rebuilt “DB version” backed by a local database entered public beta in July 2026, bringing real properties and Tana-style structured tags plus more reliable sync. That rewrite consumed several years of development attention, so the classic version evolved slowly — but the foundation of free, local, open-source outlining has kept a devoted community.
What is Tana?
Tana’s big idea is the supertag: slap #meeting or #person on any outline node and it gains fields, views and behavior — your notes become a queryable database without leaving the outline. Live queries collect matching nodes anywhere in your workspace, which makes Tana feel like Roam and Notion fused together.
Since 2025 Tana has pivoted hard into AI: a meeting agent that transcribes calls without a bot, voice capture that structures itself, semantic search, and a local MCP server so AI assistants can work with your workspace. In March 2026 the company split in two — the notes tool became Tana Outliner (outliner.tana.inc, the product compared here), while the plain “Tana” name moved to a separate agentic meeting platform. Outliner pricing as of September 2026: Free with 500 AI credits a month, Plus at $8/month and Pro at $14/month (22% off billed yearly; students, academics and NGOs get 50% off). It’s cloud-based with web, desktop and capture-focused mobile apps.
Logseq vs Tana at a glance
| Logseq | Tana | |
|---|---|---|
| Price | Free (open source); optional paid sync | Plus $8/mo; Pro $14/mo (22% off yearly) — Tana Outliner, Sep 2026 |
| Free plan | Yes — everything | Yes — 500 AI credits/mo, 3 workspaces, 0.5 GB |
| Platforms | Windows, macOS, Linux, iOS, Android | Web, macOS, Windows; iOS/Android capture apps |
| Storage model | Local-first — Markdown/Org files (DB version: local database) | Cloud |
| Editing model | Outliner — blocks, references, daily notes | Outliner with supertags (typed nodes, fields, live queries) |
| Backlinks & wiki links | Yes — bidirectional, block-level | Yes — references and backlinks |
| Graph view | Yes — graph view | No traditional graph view |
| Search | Full-text and queries | Full-text, structured queries, semantic AI search |
| Web capture | Community clippers; mobile quick capture | Mobile capture apps, voice capture |
| Email-in capture | No | No |
| RSS / feeds | No | No |
| Team collaboration | No | Yes — real-time shared workspaces |
| Publishing & sharing | Static-site export via community tools | Shared/published nodes |
| AI assistant access | Community plugins and MCP servers | Deep native AI (credit-metered) + local MCP server in the desktop app |
| Offline access | Yes — fully offline | Partial — desktop app, personal workspaces |
| Open source | Yes — AGPL-3.0 | No |
| End-to-end encryption | Yes — Logseq Sync is end-to-end encrypted | No |
| Data export | Markdown/Org files; OPML, JSON | Markdown or JSON workspace export (supertags flatten) |
Pricing and features as of mid-2026 — check each product’s site for current details.
New to personal knowledge management? Start here: choosing a second brain app — every tool on this page, with a pick for each kind of thinker.
Logseq vs Tana comparison: two outliners, one idea
Both descend from Roam’s idea: the day’s journal as the front door, every line a bullet, any bullet referenceable anywhere. Logseq kept that idea free, open source (AGPL-3.0) and local — classic graphs are Markdown or Org files on your device, and the 2.0 database version moves them into a local database with Markdown export. Tana took the same outline into the cloud and made structure the product: a #meeting supertag gives a bullet fields, views and behaviour, and live queries collect matching nodes from anywhere in the workspace.
That is the fork. Logseq asks you to own the system — sync, backups, the plugin you want — and gives you the license and the files in return. Tana runs the system for you, with real-time multiplayer and AI built in, and meters it. (Each against the files-first app: Logseq vs Obsidian and Obsidian vs Tana.)
Tana Outliner vs Logseq DB — supertags vs structured tags
“Tana Outliner vs Logseq DB” is the right question in 2026. Logseq’s database version (2.0.1 on July 13, 2026, 2.0.2 on October 7, nightlies most weeks) rebuilt the app around a local database so that tags can carry typed properties — a #book with an author and a status, queryable as a table — which is the feature set Tana’s supertags pioneered. For a solo user whose structure needs are “tags with a few fields and a table of them”, Logseq DB now does it for free.
What stays Tana-only: supertag inheritance (a #client-meeting that is also a #meeting), fields that point at other nodes, live queries you build without a query language, and a finished, hosted app. The honest flip side: Logseq DB is explicitly beta — the team recommends automated backups, its sync is alpha, the rebuilt mobile app is iOS-first — and classic Logseq is maintenance-only. Tana’s supertags were stable years ago; Logseq’s tags-with-properties are the thing being built this year.
Price and ownership — free files or a metered cloud
Tana Outliner as of November 7, 2026: Free (500 AI credits a month, 3 workspaces, 0.5 GB of uploads), Plus at $8/month (2,000 credits, unlimited workspaces), Pro at $14/month (5,000 credits), 22% less billed yearly, with a 50% discount for students, academics and NGOs (Plus $5, Pro $9 a month). Since March 2026 that is the notes product — the plain “Tana” name belongs to a separate meeting platform, whose $20–30 “Pro” is what older comparisons quote. Everything AI draws on credits, and the free tier cannot buy more.
Logseq is free: the app, every platform, no note limits, AGPL. Money enters only for sync — DIY is free because classic graphs are files, official end-to-end-encrypted sync has been reached by backing the project, and the DB version’s new sync is alpha with pricing unannounced. Logseq’s own pricing — what Sync costs, what backing buys, where the DB version stands — is broken down and dated on our Logseq vs Capacities page; we don’t restate it here. Ownership is the other half: Logseq’s graph is on your device; Tana’s is on Tana’s servers — exportable, not end-to-end encrypted, not self-hostable.
AI and agents — MCP on both sides, differently
Tana is AI-native by design — a meeting agent that transcribes calls without a bot, a voice agent that captures and files as you talk, AI chat shaped by your supertag schema — all metered by credits. For outside assistants, Tana Outliner ships a local MCP server inside the desktop app: on by default whenever the app is open, at localhost:8262, authenticated by OAuth or a personal API token, free on every plan, and able to search, read and edit nodes and their supertags from Claude Code, Codex, Gemini CLI, Cursor or VS Code. A hosted remote server exists in beta for connections that don’t need the desktop app running.
Logseq has no built-in AI and no official MCP server; the community provides both — plugins that call whatever model you like, and MCP servers over the local graph or, for classic graphs, plain filesystem servers over the Markdown folder. The difference is AI as a metered product feature versus AI you bring to your own files. What neither offers is a scope: an assistant that can reach the graph can reach all of it. (A hosted notes app with a per-connection boundary is a different design — our MCP server page explains that one.)
Switching and export
Tana exports a whole workspace to Markdown or JSON — Markdown “intended to be used for export to other tools”, JSON either as Tana Intermediate Format 2.1 (importable into a new Tana workspace) or the full stored representation. Supertags and fields arrive in Markdown as text, so a Tana → Logseq move is: export Markdown, open the folder as a classic Logseq graph (or import it into the 2.0 database version), then rebuild tags and properties by hand; node references lose their targets.
Logseq → Tana runs through Markdown too: text and nesting come through, while ((block references)), key:: properties, queries and TODO keywords arrive as plain text to re-tag. Trial-import a copy either way. If the other structured cloud apps are on your list, we’ve set Tana against them in Tana vs Capacities and Tana vs Anytype.
Where Logseq shines
- Free and open source (AGPL) with local-first storage — the strongest ownership story among outliners.
- Built-in spaced-repetition flashcards and PDF annotation, no plugins needed.
- Block references, queries and whiteboards cover most of Roam’s power features.
- The new DB version adds structured properties and tags with trustworthy sync.
Worth knowing: No real-time collaboration — it’s a single-player tool. The MD-to-DB transition splits the product; some features (like whiteboards) are temporarily missing from the DB beta. Polish and mobile experience trail the paid competition, and web capture relies on community tooling.
Where Tana shines
- Supertags are the most elegant structure system in the category — schema, views and queries grow naturally out of plain outlining.
- The AI feature set is the deepest here: botless meeting transcription, voice capture, agents, and semantic search.
- A local MCP server inside the desktop app makes it one of the few tools built for AI-assistant workflows.
- Real-time multiplayer workspaces work well for teams.
Worth knowing: AI usage is credit-metered on every plan and the free tier is thin (500 AI credits, 3 workspaces, 0.5 GB); serious use means Plus at $8/month or Pro at $14/month. Cloud-only and proprietary: no local files, no end-to-end encryption; export is Markdown or JSON, and supertags and fields flatten on the way out. The structure-first approach has a real learning curve; casual note-takers may find it over-engineered.
Which should you choose?
Choose Logseq if:
- You want Roam-style outlining without the subscription, as open-source software.
- Local files and end-to-end-encrypted sync are non-negotiable.
- You want flashcards built into the same tool as your notes.
Choose Tana if:
- You want notes with fields, views and queries — a database that emerges from writing.
- Meeting transcription and voice capture are core to your workflow.
- You’ll actually use enough AI to justify the price.
A third option: Dexi
Logseq and Tana both start from a blank outline you type into, and neither takes a forwarded email or an RSS feed. If much of what you keep arrives that way — articles, newsletters, feeds — Dexi starts there: everything you capture from your browser, inbox or RSS becomes a searchable note beside the ones you write; #tags, @mentions and [[wiki links]] typed inline do the structuring; inline key:: value properties feed table and card views when you want them; and its MCP server is hosted, with a per-connection boundary — everything, one folder, one tag — chosen when you connect, no desktop app running and no credits. Web-based, nothing to sync, free for up to 1,000 notes and $5/month after — below Tana Outliner’s cheapest paid plan.
Dexi in detail
- Web clipper and bookmark saver for any browser or phone
- Semantic search for notes: find ideas by meaning
- Knowledge graph for your notes
- Use Claude and ChatGPT with your notes (MCP server)
Frequently asked questions
Is Tana better than Logseq?
Better at structure, AI and teams: supertags with inheritance and live queries, built-in meeting and voice agents, real-time shared workspaces, and a finished hosted app. Logseq is better at cost, license and ownership: free, AGPL-3.0, your graph on your own device, and its 2.0 database version now covers tags with typed properties. In 2026 the practical split is Tana if you will pay $8–14 a month for the system, Logseq if you want to own it and can live with a beta.
Does Logseq have supertags?
Not by that name, but the 2.0 database version (beta since July 2026, 2.0.2 in October) added what supertags are for: tags that carry typed properties, so a #book can have an author and a status and be queried as a table. It lacks Tana’s supertag inheritance, node-referencing fields and query builder, and the classic file-based Logseq has only plain tags and page properties.
Is Tana Outliner free?
There is a free plan: 500 AI credits a month, up to 3 workspaces and 0.5 GB of uploads, with the local MCP server included. Serious use means Plus at $8/month (2,000 credits) or Pro at $14/month (5,000 credits), 22% less billed yearly; students, academics and NGOs pay half. Prices as of November 2026. Logseq is free outright.
Can I move my notes from Tana to Logseq?
Yes, via Markdown. Export the Tana workspace as Markdown (a zip of .md files), open the folder as a classic Logseq graph — or import it into the 2.0 database version — and rebuild structure by hand: supertags and their fields arrive as text, and node references lose their targets. Trial-import a copy first.
Which is better for AI assistants, Logseq or Tana?
Tana, as of November 2026: it ships an official local MCP server in its desktop app (on by default, OAuth or an API token, free on every plan) plus a hosted server in beta, and its own meeting, voice and chat agents. Logseq has no official MCP server or built-in AI; community plugins and MCP servers over the local graph fill the gap. Neither scopes an assistant to part of your notes.
Collect, connect and search everything you read — free, no credit card required.
Try Dexi free