Leaving Logseq? A calm guide to your options

You don’t have to decide today. Logseq still opens, your graph is still there, and the sky isn’t falling. But if you’ve been unsettled watching the file-to-database rewrite drag on, here’s the honest version: you have three options, none of them urgent, and the right one depends on what you actually use Logseq for.

Stay on Logseq OG (the Markdown version) if it does the job and you don’t want to touch anything. Wait for the database version to stabilize if backlinks and queries are the whole point of your workflow. Or move now, calmly, if the uncertainty is costing you more than the tool is giving back. This post walks each path, then shows what a move to a keyboard-first outliner actually feels like.

What’s actually going on with Logseq

Logseq spent years as a local-first, Markdown-backed outliner that PKM people loved: daily journals, block references, backlinks, and queries over your own graph, all stored as plain files you owned. Then the team started a rewrite from those Markdown files to a database backend, aiming for speed and features the file format couldn’t support.

That rewrite has taken a long time. As of 2026 the project is effectively split in two: Logseq OG, the original file-based app, and a newer database-backed version that’s been in beta rather than a settled stable release. Both work. Neither is clearly the one to bet the next five years on. That’s the whole source of the unease, and it’s a reasonable thing to feel. You picked a tool partly because it was plain files you controlled, and now you’re being asked to pick a side during a transition the team is still working through.

None of that means Logseq is dying or that your notes are at risk. It means the ground is less settled than it was, and settled is worth a lot when the tool is where you keep your thinking.

Option one: stay put

If Logseq OG does what you need and you rarely think about it, staying is a real choice, not a cop-out. The Markdown files are yours. They’re portable. Nothing forces you onto the database version, and a stable, boring tool you don’t have to think about is exactly what a note system should be.

Stay if: your graph is mature, your workflow is muscle memory, and the rewrite is a headline you scroll past rather than a problem you feel. Don’t let migration FOMO push you off something that’s working.

Option two: wait for the database version

If the reason you use Logseq is the networked, query-heavy PKM workflow, and the database version promises to do that better, waiting is defensible. Follow the beta. Try it on a copy of your graph, not your live one. Let other people hit the sharp edges first.

Wait if: backlinks and queries are the point, you’re patient, and you’d rather ride out the transition than relearn a new tool. Just be honest with yourself about how long you’re willing to wait, and set a date to re-evaluate instead of drifting.

Option three: move now, without drama

If the uncertainty is nagging at you every time you open the app, that’s information. A tool you keep second-guessing is a tool that’s costing you focus. Moving doesn’t have to be a big migration project, and it doesn’t have to be all-or-nothing. Your Markdown export isn’t going anywhere; you can start fresh in something new for daily capture and bring the old graph over gradually, or not at all.

The question is where to land. A few honest reads:

• Want the same networked-thought model elsewhere? Obsidian is the obvious harbor for ex-Logseq users who want backlinks, local Markdown, and a huge plugin ecosystem. It’s a document editor with outlining bolted on rather than an outliner at heart, but for graph-style PKM it’s the safe, settled choice.
• Want the outliner feel, kept simple? This is where a lot of Logseq refugees actually land once they admit they used maybe a fifth of the queries. If what you loved was the outline itself, fast nesting, everything in one tree, you may not need the whole PKM machine.
• Want power and don’t mind complexity? Tana gives you structured fields and supertags, but it’s a steeper climb and a paid one. That’s trading one kind of heaviness for another.

If you’re in that middle group, here’s what a lighter outliner looks like.

What moving to listlis feels like

listlis is a keyboard-first outliner built on the one idea Logseq got right: your workspace is a single tree. Each node is a title plus an optional Markdown body, so a line stays scannable until you open it up and give it room. Mark any node a todo and complete it; completion carries the whole subtree and rolls up to the root, so a finished branch reads as finished.

There’s no rewrite in progress and no version to pick. The whole loop runs from the home row. Press o to add a sibling below the current line. Press Tab to nest it under the row above, or Shift-Tab to pull it back out. Type a title; when a line needs more than a title, open a body and write Markdown. Press ! to make a node a todo, then c to complete it. Reordering is cheap because sibling order is a fractional index between neighbors, so reparenting a whole subtree writes a single row instead of renumbering the list. Press / to search every node by title and body.

What it deliberately doesn’t have is the query language. If your Logseq life was built on {{query}} blocks and advanced datalog, listlis won’t reproduce that, and you should stay somewhere that does. That’s the honest fit line. But most people don’t run a query engine. They keep an outline, capture into it, and want to find things fast, and for that the query machinery was always more than the job needed.

The part that surprises Logseq people

Here’s what tends to win over the PKM crowd once they try it. listlis has a connected assistant layer, so you can point an AI assistant at your own tree and work with it in plain language. Ask it to pull everything about a topic and it gathers the matching branches into one bundle for you, no query syntax to write. Ask what’s due and it hands back a real agenda, bucketed by overdue, today, and the next week. Capture a thought by forwarding an email, and it lands as a node in your tree. You can hand a read-only branch to someone as a link, or co-edit a subtree live.

That’s the thing queries were really for: getting the right slice of your own thinking back out when you need it. listlis just lets you ask for it in words instead of building the query yourself. Same job, less machinery.

A calm way to decide

Run this quick gut check. If you mostly write and nest and search, and you rarely wrote a real query, a lighter outliner will feel like relief, not loss. If your graph runs on backlinks and queries, either wait out Logseq’s transition or move to something that keeps that model. And if you’re genuinely happy, stay; a working system is worth more than a new one.

Whatever you pick, you’re not trapped. Your notes are Markdown, they’re yours, and the good tools all import and export it. That portability is the reason this decision is lower-stakes than the anxiety makes it feel.

If a fast, keyboard-driven outline with tasks and notes in one tree sounds like the relief you’re after, open a root at app.listlis.com and start typing. No import wizard, no template gallery, no version to choose. The first line teaches you the app.

Still weighing the whole field? The outliner comparison hub lays out where each tool fits side by side, and the Workflowy alternatives guide covers the same terrain from a different starting point. The bet underneath all of it is the one-tree idea: your thinking doesn’t split neatly into separate apps, so one tree keeps it together.