How I'm Using Logseq in 2025 - The knowledge management part of my Second Brain

Logseq remains the core of a “second brain” setup because it’s treated as knowledge management software—not a full project-management system—and its design choices make high-density reading, fast linking, and long-form writing feel natural. After nearly five years of use, the workflow sticks around despite the existence of many alternatives, largely because Logseq supports a clear separation of roles: Logseq holds ideas, references, and writing; Tana handles project management tasks like calendars and project tracking. That division reduces the urge to force complex project workflows into Logseq’s structure, while still allowing cross-search and quick navigation between tools.

A major reason for staying is how Logseq handles reference material in an outline format. The workflow favors pages and blocks differently: information can stay as blocks for quick scanning, but it can be elevated to pages when deeper context is needed. References appear in a scrollable, reverse-chronological outline with link references visible at a glance, which the user contrasts with systems where imported content hides behind extra clicks. Filtering link references is also positioned as a practical advantage—quickly surfacing related highlights without drilling into separate pages.

Editing and writing mechanics further reinforce the choice. Markdown editing keeps notes portable for long-term use, and Logseq’s block-based editing supports drag-and-drop rearranging for long-form writing. The user also values the ability to distinguish what’s original thinking versus quoted material by using editing workflows (e.g., opening content in VS Code, replacing text, and then tracking attribution). Even small UI behaviors—like an optional non-logical outliner indentation mode—are framed as meaningful because they match how the user prefers to write and reorganize.

Still, the setup isn’t without friction. Mobile performance is described as slow and cumbersome for quick capture and search, especially on Android. Calendar integration is also called out as not robust “out of the box,” with workarounds like plugins. UI limitations occasionally slow down navigation—for example, opening a linked item doesn’t always behave as expected (such as opening as a page when that’s the desired outcome).

Questions about Logseq’s future come up because releases have slowed: the last beta was April 23, 2024, with a long gap afterward. Activity exists in issue closures and development work, but the focus is shifting to “Logseq DB,” an object-based approach inspired by Tana’s super tags. In this new direction, the traditional page-vs-block distinction may change, and tags/objects would be handled differently, with Logseq DB described as using SQLite and aiming to work with markdown files. A community manager (Ramsey) is cited as saying Logseq DB is far from finished and no timeline is committed, so current tutorials could become outdated.

To adapt, the user also tweaks adjacent workflows. Omnivore is replaced by Readwise for article capture, but imports require manual restructuring so highlights become blocks visible in the outline. Over time, the user reduces reliance on namespaces, removes scratchpad-style organization in favor of MOCs, and drops “status to process” and “observation” pressure—preferring flexible tagging (including fleeting tags) and iterative loops of thought. The takeaway is less about finding a perfect app and more about matching tools to the kind of work—knowledge management versus complex project management—and then mastering the workflow until it becomes second nature.