Logseq

Overview

Logseq is a privacy-first, open-source knowledge base and outliner for personal knowledge management (PKM), research, and task work. Notes live on your device as a local graph—classically plain Markdown or Org-mode files; the newer DB (database) version stores a SQLite graph with optional Markdown Mirror export. The project is licensed AGPL-3.0, developed in the open at github.com/logseq/logseq (~44k stars, ~2.7k forks as of mid-2026), with docs at docs.logseq.com and the product site at logseq.com.

Primary job: capture daily thoughts in a journal-first, block-based outline, then grow a linked knowledge graph with bidirectional references, queries, PDF annotation, whiteboards, flashcards, and a plugin marketplace—without locking your data into a proprietary cloud format. Desktop apps ship for macOS, Windows, and Linux; mobile apps exist for iOS and Android; a web client is available at app.logseq.com for the DB path.

In 2025–2026 Logseq explicitly split into two product lines: the long-running file-graph app (stable-ish releases such as 0.10.15, December 2025) and the Logseq 2.0 / DB beta (e.g. desktop 2.0.1, July 2026) with RTC-style sync, richer properties/tags, CLI, and MCP tooling. Official roadmap and the “two versions” announcement are published as public Logseq graphs on logseq.io. Core local software remains free; optional hosted sync/publish are paid or backer-gated via Open Collective.

Key features

• Local-first graphs — File graphs: plain Markdown/Org on disk under your control. DB graphs: SQLite with graph export (SQLite, EDN, standard Markdown) and opt-in Markdown Mirror so you can still grep/git a projection of blocks.
• Block-based outliner — Every bullet is a first-class block with indent hierarchy, zoom-in focus, embeds, and references. The DB model further unifies “nodes” (pages and blocks behave more alike, including linked/unlinked references).
• Daily journals — Journal-first workflow for capture; natural-language dates and date pickers; journals are tag-backed (#Journal) in the DB version and customizable with tag properties.
• Bidirectional links & graph view — Page/block links with backlinks; Graph View rebuilt as V2 in 2026 DB releases for better performance on large graphs and task neighborhood zoom.
• Queries — Simple queries plus advanced Datalog-style queries (DataScript heritage). Query Builder for guided construction; DB version exposes query tables/views and CLI querying (e.g. qmd query).
• Tasks & properties — TODO workflows, priorities, deadlines/scheduled dates, recurring tasks (repeat cookies .+ / ++ / + semantics fixed in DB). DB graphs add typed properties (text, number, date, checkbox, URL, node, asset), tag inheritance, and spreadsheet-like table views.
• PDF annotation — Highlight and annotate PDFs inside the app; annotations become graph-linked blocks (file + DB paths).
• Whiteboards — Free visual canvases for diagrams and spatial thinking (widely associated with the Excalidraw-class drawing stack); shipped free to all users rather than as a paid exclusive.
• Flashcards / spaced repetition — Built-in cards; DB reimplementation uses the open FSRS scheduler.
• Plugins & themes — Marketplace of plugins/themes; Plugin API at plugins-doc.logseq.com. File-graph ecosystem is large (hundreds of plugins historically); DB support is smaller but growing (docs cite 65+ DB-capable plugins mid-cycle) with a distinct SDK surface.
• Publish — Share pages/graphs to the web; DB Publish targets logseq.io with optional passwords and self-host URL config (Cloudflare-oriented). Older full-graph publishing remains documented for classic workflows.
• Sync & RTC — Official encrypted file Sync (beta historically via Open Collective). DB path adds RTC (real-time collaboration) sync—invite/backer gated in docs, with self-host Sync Server URL settings.
• CLI, scripting, MCP — Bundled CLI for graph ops and automation; nbb-logseq for scripted DB read/write; optional MCP server so local LLM tools can create/search nodes with undo support in the app.
• Cross-platform — Electron desktop; Capacitor-based mobile (DB iOS alpha/invite; Android native UI in progress per docs). Linux install script available from the repo.

Pricing

Logseq’s commercial model is open-core / freemium on hosted services, not seat licenses for the editor. The local application is free and AGPL-licensed. Money funds infrastructure and team work mainly through Open Collective donations and eventual “Logseq Pro” hosted offerings.

Tier / path	Price (USD)	What you get	Notes
Local app (file or DB graph)	$0	Full outliner, queries, plugins, PDF, whiteboards, offline use	AGPL-3.0; no seat fee
Open Collective Backer	$5 / month	Supports development; historically required for Logseq Sync beta access	Recurring monthly on opencollective.com/logseq
Open Collective Sponsor / Supporter	$15 / month	Higher support tier; early Sync/DB testing callouts in 2026 announcements	Same Open Collective page; higher sponsor tiers ($100+, $250+) also exist
DIY sync	$0 (+ your cloud)	Git, Syncthing, iCloud/Drive/Dropbox on the graph folder	Common free alternative; conflict risk if two devices write without care
Self-host Sync / Publish	Infra cost only	Community guides for hosting Sync and Publish backends	Settings allow custom Sync/Publish server URLs in DB docs
Logseq Pro (stated direction)	TBA (hosted only)	Sync, real-time collaboration, publishing—not local feature locks	Core team clarification on forum: Pro = services that cost servers to run

What is free forever in practice: running Logseq offline on your machines, storing graphs locally, using the plugin marketplace, and exporting/backing up data. You do not need an account for basic desktop file-graph use.

What costs money: convenience and multi-device polish via official Sync/RTC, and official hosted publish when you use Logseq’s infrastructure. Exact general-public subscription SKUs beyond Open Collective tiers have long been “coming”; treat $5/$15 as the real numbers people pay today for backer Sync, not as a final SaaS price card.

Limits & gotchas

• Two-product confusion — File 0.10.x and DB 2.0.x are not the same UX or plugin surface. Migrating file → DB uses an importer with known limitations (e.g. only one query/embed/quote per block when the file graph had several). Plan a test graph.
• Learning curve — Block-first outlining, Datalog queries, and (in DB) tags-as-types are powerful but steeper than a simple notes app. Casual note-takers often bounce to simpler tools.
• Development pace perception — Reddit/HN threads frequently call the file app “stale” while energy goes into DB. That split frustrates users who wanted incremental file-graph polish instead of a storage rewrite.
• Mobile reliability — Community repeatedly cites mobile bugs, sync duplicates, and missing parity vs desktop as the reason they stay on Obsidian or single-device Logseq. DB iOS is invite/alpha; Android DB app not fully opened for alpha in the same docs snapshot.
• Sync conflicts & data trust — HN and forums report occasional “I know I wrote that” data-loss feelings on file graphs; DB is partly motivated by stronger consistency, but beta status means backups are mandatory.
• Performance at scale — Large graphs can feel heavy (Electron + rich indexing). Graph View V2 and DB storage aim to improve this; huge media-heavy graphs still need discipline (assets folder, pruning plugins).
• Plugin fragmentation — Many popular file plugins do not yet “Support DB.” Check the marketplace DB badge before committing a workflow to 2.0.
• Electron security surface — Desktop is Electron-based; community comments (including on the 2.0 HN thread) criticize dependency freshness. Prefer official releases, keep OS auto-update, and treat third-party plugins as code you trust.
• AGPL implications — Fine for personal use and most self-host. Organizations modifying and distributing network services should read AGPL obligations with counsel—not a “do anything MIT” license.
• Not a team Notion replacement out of the box — RTC collaboration is the strategic answer, but it is gated/early; multiplayer knowledge bases still often choose Notion, Affinity-class tools, or self-hosted stacks.

Community sentiment

Sentiment is polarized but stable: power users who live in outliners and local files love Logseq; people who need polished mobile sync often leave.

Praise (Reddit, forum, HN, Product Hunt): block-level linking beats page-only wikis for atomic notes; daily journal + tasks + PDF workflow feels like “org-mode for the rest of us”; privacy and plain-text longevity; whiteboards and query features free for everyone; DB adopters report properties/tags/tables as a real step toward personal databases without leaving the outliner.

Criticism: slow visible progress on the classic app; mobile quality; occasional sync/duplication bugs; fear of monetization after funding; risk and migration cost of DB; Electron weight; advanced queries hard for non-programmers. Threads titled along the lines of “is Logseq dead / stagnating / what will you replace it with” recur even as DB releases land—signaling communication and packaging issues as much as raw capability.

App Store / Play reviews and forum posts echo the same split: desktop enthusiasts vs multi-device frustration. Product Hunt historically celebrated the joyful private digital-garden angle; longer-term PKM communities compare Logseq constantly to Obsidian.

Who should use it

• Researchers, students, and Zettelkasten-style writers who think in bullets, PDFs, and linked claims rather than folders of long pages.
• Developers and Emacs/org-mode alumni who want Datalog queries, CLI/MCP automation, and git-friendly (or SQLite-exportable) graphs.
• Privacy-sensitive individuals who refuse cloud-only notes and accept managing backups themselves.
• Solo knowledge workers happy on one primary desktop with optional phone capture—not necessarily full mobile editing parity.
• Early adopters willing to run DB beta, file issues on db-test, and keep hourly backups for the new property/tag model.

Who should skip (for now): teams needing mature real-time multiplayer today; users who want “install and forget” phone sync like Apple Notes; people who only need short lists (overkill); orgs that cannot run AGPL evaluation or Electron desktop apps.

Alternatives

• Obsidian — Local Markdown, huge plugin ecosystem, polished mobile; page-centric (not block-outliner-first). Sync ~$4/mo, Publish ~$8/mo. Best when you want stability and marketplace breadth.
• Notion — Cloud databases, teams, templates; weaker offline/privacy story. Choose for collaboration and structured tables without self-hosting.
• Roam Research — The commercial block-graph pioneer Logseq is often compared to; paid cloud, less local-first.
• Anytype / Capacities / Reflect / Tana — Object- or graph-oriented PKM with different sync and pricing models; evaluate if you want modern UX with less DIY.
• Org-mode (Emacs) — Maximum power and longevity if you already live in Emacs; steeper environment cost.
• Joplin / Trilium / SiYuan / AFFiNE — Other open or local-first note systems; pick based on outlining vs wiki vs all-in-one canvas needs.

Verdict

Logseq remains one of the strongest free, open-source choices for a local, block-based second brain—especially if you value AGPL transparency, journals-plus-outliner workflow, PDF annotation, and deep queries. The honest 2026 picture is not “finished polished SaaS,” but a dual-track project: a mature file graph many still run daily, and an ambitious DB 2.0 beta that adds properties, tables, RTC, CLI, and MCP at the cost of migration and stability risk.

Pay $0 if local-only or DIY sync is enough. Pay about $5–15/month via Open Collective if you want official Sync/backer access while Pro packaging finishes. Prefer Obsidian if mobile polish and plugin abundance matter more than block-outliner semantics; prefer Notion if the team is the product. For individuals who will invest in the model, Logseq’s combination of ownership and outlining depth is still hard to beat on price.