Skip to content
AgentThread
Open Source#note-taking#knowledge-management#local-first#open-source#privacy#pkm

Logseq: A Local-First Note-Taking Tool That Connects Your Ideas Automatically

An open-source outliner and knowledge base for researchers, analysts, and knowledge workers who want their notes to form a network, not a folder tree, with files that stay on their own device.

AgentThread7 min read
Share

If you take a lot of notes and find yourself forgetting you wrote something six months ago, the problem is usually organization. Folder-based systems force you to decide upfront where something belongs. Logseq takes a different approach: write freely, and let the tool surface connections automatically. Notes that mention the same topic link to each other without you doing anything. Over time, a picture of your thinking builds itself.

Logseq is not trying to be Notion, the hosted all-in-one workspace product. It is a focused tool for people who think in writing and want their notes to be useful again in a year, not locked in someone else's database. The project lives at github.com/logseq/logseq and has 44,000+ stars.

FactValue
What it isA local-first, open-source outliner and knowledge base with automatic note connections
Built byThe Logseq open-source team, funded through Open Collective
LicenseGNU AGPL v3 (free to use; open-source with copyleft conditions)
PlatformsWindows, macOS, Linux desktop; iOS mobile; web app at app.logseq.com
RequiresNo account for desktop use; no internet connection required
Install methodDownload from github.com/logseq/logseq/releases
VerdictThe right tool for a knowledge worker who wants a private, file-based thinking system that connects ideas without manual work

What Logseq actually is

Logseq is an outliner (a note editor where every line is a bullet that can be collapsed or expanded, so you build a hierarchy of thoughts) combined with a backlink system (automatic connections Logseq creates between notes that mention the same topic).

The two features together produce something different from a regular note app. In a regular app, a note about "project kickoff" and a note about "Sarah's feedback on the proposal" are two separate files unless you manually link them. In Logseq, if both mention "Q4 roadmap," Logseq automatically surfaces each on the "Q4 roadmap" page as a linked reference. The connection forms because both notes used the same phrase, not because you filed them together.

Logseq is local-first (your notes live as plain files on your own device, not on a company's server). The file format is Markdown (a simple text formatting system using symbols like # for headings and * for bullets). Because notes are plain files, they are readable by any text editor, portable between tools, and not subject to format changes by the vendor.

The graph view (a visual map of how your notes connect to each other) makes the web of connections visible. It is useful for finding clusters of related thinking and for spotting isolated notes that have never been mentioned anywhere else.

The project was first released in 2020. Version 2.0 is in beta as of mid-2026.

What you get after installing it

The daily journal page is where most users start each day. The app opens to today's dated page. You write bullets. If you wrap a project name, person's name, or concept in double brackets, Logseq creates a page for it and registers your entry as a linked reference there. The journal is the input buffer; the pages become the permanent record.

Backlinks and linked references are the core mechanic. Every page shows every other note in your database that mentions it. A page called "competitive analysis" accumulates every note you ever wrote that mentioned it, across months and projects, without any filing effort on your part.

Queries let you filter notes by properties, tags, dates, and task status. A query that shows every note tagged "action-item" not yet marked done is a two-minute setup. Tasks live inside notes and surface across your database on demand.

PDF annotation is built in. You can open a PDF inside Logseq, highlight text, and the highlighted excerpt lands in your notes as a linked reference back to the exact page of the source PDF. Researchers who work from documents find this the most useful feature in practice.

Plugins extend Logseq through a JavaScript API. There are plugins for Zotero (an academic citation manager), Anki (a flashcard review tool that drills you on material at increasing intervals until it sticks), calendar views, and many others. The marketplace is built into the app.

The install experience

The install is a file download. Go to github.com/logseq/logseq/releases, download the installer for your operating system (macOS .dmg, Windows .exe, or Linux .AppImage, a self-contained Linux program file), and run it. No account, no API key, no server setup. The app opens to your first daily journal page in under two minutes.

First setup involves picking a folder on your computer where notes will live. Logseq reads and writes .md files in that folder. If you already have Markdown notes from another tool, point Logseq at that folder and it indexes them immediately.

Syncing across devices works through any folder sync tool you already have: iCloud, Dropbox, or Syncthing (a free, open-source tool that syncs folders directly between your own devices with no cloud intermediary). Logseq does not require its own sync service. Mobile is iOS for the current stable version.

The one honest friction point: onboarding does not hand-hold. You land in the journal and write. The backlink model clicks faster in practice than it looks like it will from reading about it.

Where it fits and what to compare it to

Notion is the most common comparison. Notion is better for team wikis, project management boards, and anything requiring real-time collaboration. Logseq is better for personal knowledge management where data privacy matters and you want notes to last beyond the vendor. Your Notion data lives on Notion's servers; Logseq data lives on your drive.

Obsidian is the closest direct competitor. Both are local-first Markdown tools with backlinks and graph views. Obsidian's graph view is more polished and its plugin ecosystem is larger. Logseq's daily journal and outliner model is better for users who think in bullets and want frictionless daily capture. Both are strong choices and the files are compatible, so switching between them is low-risk.

Roam Research (at roamresearch.com) invented much of the mental model Logseq is built on: the daily journal, bidirectional links, and the block-based outliner. Roam is hosted (not local-first), proprietary, and costs $15 per month. Logseq is the open-source equivalent for users who want the same model without the subscription or the data privacy tradeoff.

The specific cases where Logseq is the right answer: you read a lot and want your reading notes to connect; you work under confidentiality constraints and cannot put notes in a hosted tool; or you have tried folder-based systems and found that notes disappear into the filing cabinet and never come back.

Verdict

Install it if you are a researcher, analyst, writer, or professional who takes notes regularly and wants them to form a usable knowledge base rather than an archive. The backlink model and daily journal habit are genuinely different from folder-based note-taking, and the difference compounds over months.

The caveats: Logseq does not do team collaboration well in the current stable version. The database version (in beta as of mid-2026, storing notes in SQLite rather than plain files) adds real-time sync and multi-device support, but carries data-loss risk and should not be used for anything critical yet. The plugin ecosystem is smaller than Obsidian's. The outliner model is a strong opinion: users who prefer writing in prose rather than bullets often find it limiting.

For a user who fits the profile, none of those caveats are disqualifying. The AGPL license means the source is open and auditable. The file-based version has been stable for years. The project has 44,000+ GitHub stars, an active Discord community, and is in active development. Two weeks of actual use is enough to know if it fits.

Related posts