How I use Tags vs Links in LogSeq

Logseq’s “tags vs links” distinction is mostly visual, but the practical choice matters for how information gets organized, filtered, and surfaced. Bracket links are embedded in the writing flow—often inside sentences—while hashtags function more like labeled markers that you place at the end of blocks. That difference drives most of the workflow decisions: use tags sparingly as bundle-and-filter primitives, and default to links when you want readable, sentence-like references that can include multi-word titles.

Tags in Logseq behave like colored labels in a book: they’re best for grouping related ideas and for quick filtering later. They’re typically written as one-word hashtags (multi-word hashtags get treated like bracket links with a hash prefix), and themes may render them with distinct colors. The transcript notes that there’s no major data-level difference between tags and links, but Logseq treats them differently in edge cases—such as generating a nice list when using page properties and powering the Tag Manager plugin, which shows only tags and ignores links.

Several concrete tag use cases are offered. When journaling, hashtags like “idea” capture thoughts that should be revisited later. In script work, tags like “green screen” act as a lightweight retrieval system: later, it’s easy to “show me all” blocks tagged with that label and jump straight to the relevant take. Because tag colors are theme-driven, customizing them requires editing the CSS theme code; the creator suggests copying an example line and changing the tag name and color.

A special tag variant is also highlighted: using the page-level `tags` property. It applies tags to an entire page rather than a specific block. The workflow has shifted over time—page properties used to support a “type” system (e.g., “type book” vs “type project”), but the plan is to move toward more explicit page tags like “book tag” and “project tag” so books and projects automatically appear as clean lists.

Links, by contrast, are positioned as the default when in doubt. They’re written as part of sentences using brackets, which makes them more readable and natural for entities like people (first and last name), projects (often multi-word titles), and books. The transcript also recommends avoiding “linking tags” or “tagging links,” since mixing the two can create confusion.

Finally, the Tag Manager plugin is used to keep the tag system tidy. After installing and activating it, a tag list appears with search and sorting options based on tag name or usage count. The workflow includes merging duplicates (e.g., combining “tag” into “tags”) and notes that merges may require refreshing and re-indexing so usage counts update correctly. The takeaway is that tags are powerful for lightweight labeling and filtering, while links are better for readable, sentence-integrated references—together forming a practical division of labor in Logseq.