What if I you had to move to Obsidian?

I recently asked myself how easy it would be to migrate to Obsidian if I had to stop using Logseq for some reason? I’m happy with Logseq, but like to know that I could switch if I had to - it gives me peace of mind about my data.

The question is really a combination of several issues: user interface, features, and data. Out-of-the-box Obsidian can’t really handle a logseq-style workflow, but there are a number of plugins that can help:

These take care of many of the user interface and features that one might worry about, such as outlining, queries, and tasks. (Not entirely, since Obsidian isn’t truly block based - but close enough that one could manage.) But what about moving your data over?

As I see it there are four areas one has to worry about when migrating: block references, block properties, page properties, and tasks. Fortunately, when you create block references Logseq creates a hidden property that can be searched for in Obsidian, so that isn’t an issue. Less fortunately, Obsidian doesn’t seem to have any concept of block properties. I don’t use those much, but for people who do that could be a problem. I don’t know if there is a work around?

Regarding the last two, page properties and tasks, it seems possible to convert to Obsidian and the Obsidian Tasks plugin using clever search and replace tricks. I haven’t tested this fully - just enough to ease my mind that it is possible. I’m not a programmer but I do know some regex (grep). I use BBEdit’s multi-file search and replace tool on my vault. Here is one example:

FIND: DOING \[#C\](.*?)$
REPLACE: [/] #task \1 🔼

I initially had difficulty because I wasn’t placing the priority marker at the end of the line, but this works. (Note: I had to add the [/] “custom status” in the settings beforehand.) I didn’t do anything similar for page properties, but that seems much simpler. Perhaps someone has already written a script to do this?

Anyway, my sense at looking at this was that it would take me one weekend of work to be up and running in Obsidian if I had to switch for some reason, and perhaps another few weeks before I was able to re-establish all my existing workflows in the new environment. Not something I would do lightly, but nice to know that it can be done.


For me what’s make Logseq unique is that page references are inherited by children, so you can retrieve blocks using Linked References filters or queries. Block properties also help a lot with queries.

This implies that one can take advantage of the outliner UI to tag many blocks at once, just as showcased by Logseq’s changelogs.

And the retrieved blocks are transcluded, so they can be edited live.

1 Like

What alex said matches exactly my analysis: obsidian has no concept of tag inheritance, which I use extensively.

I evaluated obsidian for two main reasons:

  • I wish I could query using SQL or JS, I don’t want to study another language just to query my notes, but it looks like I will have to
  • I’m concerned about Logseq roadmap, which doesn’t reflect at all what I was hoping for: improving the query language being key, as well as getting abilities to build text data based on query, kinda like decipad. That together with many improvements over query reloading

Obsidian has also way wider adoption. That being said, it just doesn’t work like Logseq, I would lose too much.


I really didn’t intend for this to be a discussion of the relative merits of the two platforms. I am a logseq user myself, and don’t disagree with the comments above. However, it gives me comfort to know that I could access my data and continue some of my existing workflows on another platform if I had to. In fact, I think this is another reason to support Logeq - because one is not locked in.


I am currently working in Logseq but in the past I had about the better part of a year that I changed between Logseq and Obsidian every month or so.

This alone proves that it is very doable to make the switch (I do not use block references in my notes only links). My biggest issue in that time is that logseq saves metadata (collapsed: true) in the note. So I had a lot of those when I made the change to Obsidian.

The other way around was fairly easy to do only the indentation is gone.

The last half year almost I have been on Logseq exclusively as the stability and speed has improved quite some although not on par with Obsidian’s.

A few things I which Logseq would add are

  • Logical linking to (sub)bullets. In Obs you can link to a page#header and I would really like if Logseq could do that also. Something like page#bulletpoint#subbulletpoint, without the need for the strange number it generates now.
  • Have a working map solution to show multiple (or all) geo-data-point on a map.
  • Nested tags → Something like #person/familie and #person/college
  • The way you can work with attachments.

I realize that pointing out differences was not the point of the tread so I will stop now :slight_smile: .


Same here!
I tried to move my main work graph to the obsidian just for fun. And used the format converter to convert Logseq’s tag into link in obsidian. It workd fine. But when I open the page to see the back links, I only see the content of the linked blocks but not their child blocks. I use the parent block as titles and store the content in the child blocks. And now I only see a list of titles. [Sigh] SO NO AUTO SUMMARY PAGES like you get in Logseq.
And you can’t edit it in the back link.

Guess I’m not migrating to obsidian in the foreseeable future.
Logseq is just so good at summarizing pieces of notes and present a summary page for you to look at it and get the change in things.

1 Like

Shit, the plugin Influx enable you to see the child blocks in the backlink.
Although you still can’t edit it.

1 Like


  • Logical linking to (sub)bullets. In Obs you can link to a page#header and I would really like if Logseq could do that also. Something like page#bulletpoint#subbulletpoint, without the need for the strange number it generates now.

I’m a regular Obisidan user and while I like being able to link to headers, if you edit the header name the link is broken (see here). Maybe that’s why the strange number is needed as I wonder if a similar number would be needed in Obsidian, the same way its referenced blocks need unique IDs.

Then again I’ve barely used Logseq and am unfamiliar, but tempted by it. I thought it didn’t have WYSIWYG but the recent videos I’ve watched make it seem like that could be wrong. If solid Vim supported got added I might make the jump.

1 Like

Maybe only ### header blocks should be linkable like that. Like pagename change also the # header change would be applied everywhere. With DB version this should be possible.

1 Like

Vim plugin is available in Logseq already, you can just have an hour and play with it.

I found this discussion very useful. Thanks to all of you for sharing your thoughts and experiences. I’ve briefly tried obsidian and it seemed workable but I wanted something that came closest to replicating roam research - that app that initially seemed to be the solution to everything but fizzled fast.

On a related note: are obsidian and Logseq the only two major apps that save data in human readable text files? Just curious if there are other options out there but it’s comforting to know that Logseq can be replaced by obsidian for now.

Thanks. I heard that the vim plugin was good but lacking more features vs. VS code or Obsidian’s implementations; does it still work well enough for you? It’s great to have it either way.

@pendolino There are many other editors that save data like YAML frontmatter in Markdown, such as https://silverbullet.md/ (unique as it runs in a web browser and sync from a server), and I guess emacs and org-mode could probably do that. Though I’m guessing you’re looking for a more specific type of data saving.

1 Like

Thanks. Silverbullet looks cool though I’m looking for a solution with cloud syncing capabilities. This would take to much work for now but still relevant.

There are also QOwnNotes and Zettlr.

1 Like