Logseq DB Unofficial FAQ

Questions & Help FAQ
1

WARNING: Logseq DB is still in alpha. Use only at own risk of losing data.

This post is for people who have questions about the new database (DB) version of Logseq. For the official announcement about the database version, see this post from April 2024.

In writing this FAQ I don’t claim to speak for the team. Also, I am not a programmer and might make some mistakes. I’m just doing the best I can based on information I’ve seen posted on Discord and the Forums.

You can link directly to the FAQ with this short URL: https://bit.ly/logseqfaq

Table of Contents

What happened to markdown files?

Both the original Markdown (MD) based version of Logseq and the DB version use an internal database when you are using the app. The difference is that, in the MD version, the files are always canonical. In the DB version, while you will be able to export to markdown files, the database version is always canonical.

This chart from @danzu shows how it works. When you are using either version of Logseq, it reads from in-memory database running on “Datascript.” The MD version stores the disk version in MD files, while the DB version stores the disk version in SQLite (an embedded database).

storageformat
storageformat1916×1254 182 KB

What does this mean? In the MD version, you can modify your graph by editing the MD files directly. In the DB version you cannot do this. Some flexibility is lost as a result, but in exchange we will get better sync, real time collaboration (RTC), and less data loss.

Note, however, that sync and RTC will be part of “Logseq Pro” and require a subscription. See more on sync below.

The team is looking into the possibility of having continuous real-time export of your database into markdown files, but haven’t committed to doing so. As I understand it, this would be one-way sync, and would not read changes in the markdown files back to the database.

Can I continue using the MD version of Logseq?

Yes, the new desktop app supports both MD and DB graphs, and will continue to be supported. However, the new web-based version of the app will only support DB graphs. On mobile there will be separate apps for MD and DB graphs. Also, MD and DB graphs will sync in different ways, with MD graphs using the existing sync solution, while the new DB graphs will require RTC sync, which will be part of “Logseq Pro.” (More on sync below.)

What are NewTags?

Actually, they are just called “tags” but I call them “NewTags” here to make it clear that things have radically changed from the current MD app. This is one of the most exciting features of the DB version and will radically change how you use the app. The short version is that properties are now associated with NewTags, and assigning a NewTag to an item will automatically populate that item with those property fields.

So, for instance, you could associate “Telephone Number” and “E-mail Address” with #Contact .

See my post on how tags work in Logseq DB for more information.

How (and why) to extend tags?

Even better, you can build on existing NewTags by creating tags that inherit the properties of their parent tags. This is called “extending tags.” How this works is explained in this post by @brerlapn.

Can we import our MD graph?

Yes, you can. Right now it is recommended to use the “File to DB Graph” option, and not the “EDN to DB Graph” option.

Note that #tags and [[page references]] are handled the same in the MD version but are different in the DB version. (See about NewTags above.) This will affect how your graph gets imported, depending on which options you choose during the import process. You should experiment with different options to see how you want to handle it. Importing will not affect the original MD files.

Personally, I recommend playing with the new DB graph on its own for a while first, so you understand how it is different before importing your old data.

What happened to whiteboards?

Whiteboards have been removed from the DB version of the app, temporarily, but will still work in the MD version of the combined desktop app.

They will come back after the team decides how best to handle them.

What happened to Zotero?

Zotero will no longer be supported as part of the core app. One will have to rely on plugins, like Zotero-local. The developer of Zotero-local has promised to update the app after the desktop app is released.

Is there a desktop app for Logseq DB?

A new combined desktop app is due to be released shortly. You can already try a pre-release build from this page. (You have to be logged in to GitHub in order to see the download links.)

This app allows you to open either MD or DB graphs. The DB version is still in alpha though, and they strongly recommend against using it for important work. Also, back up your MD graph before trying it as well, and note that some changes may affect how your MD graph behaves, especially if you depend on plugins. (Not all plugins will work with this new version of the app.)

The new combined (MD and DB) desktop app fixes hundreds of bugs for MD users, but also breaks some workflows, especially those related to plugins, or some features which have been removed to become plugins, like Zotero integration. (See below for more on this.)

The DB version also exists now as a web app, which still stores data locally (not in the cloud). Everyone is free to try it at https://test.logseq.com/ (It is recommended that you only use it with Google Chrome, and turn on the backup feature under the “export” menu on the top-right.)

The old, MD-only, app still exists and gets occasional updates. It can be downloaded from the Logseq homepage. That is the 0.10.x version. The new, combined app is the 0.11.x version.

What happened to sync?

There will be two sync solutions going forward. MD sync will continue to work as it does now, while the DB version will be part of “Logseq Pro” which will include real time collaboration (RTC) and publishing as well. It should be much faster and more reliable than the MD sync.

Logseq Pro will hopefully make Logseq profitable and allow for the company to be sustainable over the long term, so the team is putting a lot of work into making sure that these features work flawlessly.

Note that there will be two different mobile apps going forward. The old ones will continue to be updated and will support MD graphs, and the new ones will be for DB graphs. Each will sync in a different way.

MD graphs in desktop app ↔️ current sync solution ↔️ old mobile app

DB in desktop or on web ↔️ RTC ↔️ new mobile apps

What is Logseq Pro?

See the official FAQ.

Will we be able to self-host Logseq DB?

See this discussion

When will it be released?

Tienson shared this internal roadmap on September 25th. Since they have made some notable improvements to data integrity and the plugin API. A full changelog can be found here, and a short summary of what they have done over the past year can be found here.

image
image1970×1712 222 KB

The final, “Research,” section will probably not be part of the initial release, but almost all the other features (except for publishing and the planned calendar feature) are already there, though some still need work. Since this chart was first shared, they have already released versions of the Server API, the MCP, and the plugin API.

There is no official timeline, but the new desktop app, iOS mobile app, and RTC sync are already being used regularly by some testers. The Android app should be released soon, as well as an update to RTC that will support end-to-end encryption. Once these are deemed reliable enough, the testing pool will be expanded, and eventually opened up to anyone who wants to try.

Right now, the new web app, and the new desktop app (see above) are already available for testing by anyone who wants to try them.

Will there be a calendar view?

Yes, supposedly a calendar view is planned. There are no details on it yet.

Will the mobile app support plugins?

It seems that there are plans to do this eventually, but right now the priority is to release the app and sync, so work on this will not happen for some time.

What is the “Library”?

One feature that has gotten very little attention from Logseq over the years is the “namespace” feature, which gives users the ability to have some additional hierarchy. In the MD version this feature was never even properly documented.

It seems that, in the MD version, the primary use intended for the feature was to avoid conflict between two identically spelled words that have different meanings. For instance, you might want to differentiate a page for “apple” when applied to Apple Computers, and for Apple Pie. You could do this by putting them in separate namespaces: [[computer/apple]] and [[pie/apple]].

Now, however, the team has added a new “Library” feature to the DB app, which should make this easier, and make namespaces more useful.

What is Library? According to the PR for this feature , it is a “page to automatically collect namespace pages.” You can watch the demonstration video on that page, but the idea is simple. If you write [[computer/apple]] it will actually make it [[library/computer/apple]] and on the page [[library]] you will see an outline of all the page names that have hierarchy. You can edit them by simply dragging and dropping those pages in the outline, and you can add additional namespaces as well.

This should make editing hierarchy much easier, and allow people to give their Logseq notes more structure if they need to.

Note: this feature is still in development.

How Can I Help Test the Mobile Apps and RTC?

Here is the form to request to be included in the pool of testers for the new iOS and Android apps:

And here is the form to request to be included in the RTC (Real Time Collaboration) test pool, necessary to sync mobile apps with the desktop:

What is the MCP and how to use it?

According to Wikipedia, “a Model Context Protocol (MCP) is an open standard, open-source framework introduced by Anthropic in November 2024 to standardize the way artificial intelligence (AI) systems like large language models (LLMs) integrate and share data with external tools, systems, and data sources.” Even though it was introduced by Anthropic (the makers of Claude AI), it is now widely supported.

The Logseq MCP can interface with online LLMs like Claude or OpenAI, or offline ones, like Gemma, to let you directly interact with your graph. Like the API, this requires using the desktop app, not the web app (see above).

Documentation is still rough, but instructions for setting up the new Logseq MCP can be found here.

18 Likes
2

Thanks for explanation.
I am using the desktop app for Logseq DB.
At first, I can not get used to the new tag. Through your post, I learned that they are targeting the super tag of tana.
After reading official document, I get the point of new tag.
I believe that the release of logseq DB can put logseq on the right track. After making profits through membership services, the update speed can be accelerated.
This will attract more developers to develop plugins to enrich the entire software ecosystem.
I hope that the official synchronization of logseq will be the same as Obsidian, and Chinese people can synchronize without VPN.

3

Since the founder and lead developer, Tienson Qin, is based in China, I don’t think that will be a problem.

5

Thank you so much for your incredibly helpful post(s), @Luhmann! I’ve been a long-time Logseq user and your writings have been instrumental in allowing me to get the most out of Logseq as well as in now slowly transitioning to Logseq DB!

That said, there is one challenge I am facing with the new DB version that I haven’t found a workaround for yet, and I’m very much hoping you might be able to help with this:

I’ve been using attributes in Logseq MD quite a lot. For example, I’d have a page for [[Jon Krakauer]] and then I’d add an alias:: [[Krakauer, Jon]] for academic referencing. For a few personal contacts, I’d even add the first name [[Jon]] as an alias as well. This workflow no longer seems to be possible, so I was wondering whether there is any workaround or how you’d go about this. Thank you so much in advance for your help!

6

Hi. Steven.
There is an official property in Logseq. It is Alias.

image
image333×138 4.81 KB

2 Likes
7

Hi @Ryuki, thank you so much for pointing that out, this is absolutely amazing and it perfectly works! Gosh, I’m so relieved that I can continue with my existing workflow! Just out of interest: For those of you who have been using the DB version those past weeks, has anyone experienced any data loss or anything alike?

2 Likes
8

Actually, in my two weeks of low-intensity use, I didn’t meet this issue.
Although there is no problem with the data, I did crash the software several times during my initial exploration. So the current alpha version is not stable enough. You need to be very careful when using it.

2 Likes
9

So Logseq MD will continue to be supported on desktop and mobile and I can always decide whether I use DB or MD for new graphs, correct?

What features will not be available or not work as well when I choose an MD based graph? (e.g., sync and RTC)

What features will not be available or not work as well when I choose a DB based graph? (e.g., self hosting with sync)

11

Here is a chart from @danzu

Logseq_-_Feature_Comparison
Logseq_-_Feature_Comparison1920×5282 458 KB

7 Likes
12

Thank you @danzu! That chart is extremely helpful and makes perfect sense.
What are the longterm plans for MD mode? Is the plan to continue to support and develop MD mode further? Or will MD mode eventually be deprecated? I imagine it to be kind of challenging to support both modes well.

13

Whenever the team has been asked such questions in the past, they have insisted that they will support MD mode going forward. But it is hard to know exactly what that means. I think we got a sense from the last big update to the desktop app, which broke some plugins and themes that worked with MD mode, even as it fixed hundreds of bugs. So the “core” MD mode still works, and was even updated with bug fixes, but some tools MD users might have depended on were broken. It is also unclear whether or not they will continue to support the older version of sync used by MD users?

1 Like
15

Thanks so much @Luhmann and @danzu .

Could I ask what the Advanced Properties are though :thinking:. I thought they were just the ‘class’ feature but I guess not. Maybe I missed an update.

1 Like
16

You are truly a genius at collecting and organizing

2 Likes
17

The ‘class’ feature refers to NewTags, which allow you to associate a number of Advanced Properties with a tag, as well as have an automatic table view of all tagged items. Advanced Properties refers to the features associated with the properties themselves. You can create different kinds of properties: date, text, node, checkbox, etc. and they have a number of settings associated with them. The best way to understand it is to create a NewTag yourself and add some properties. You will then see all the options that are available.

1 Like
20

That is a little difficult to understand. Maybe we need some examples in real-world use cases.

Thanks for your post. It really helps us.

21

Take a look at Logseq DB Library video: https://www.youtube.com/watch?v=l1xpHrZKGIY
Let me know if there is anything else to clarify.

22

My workplace has limited network access. I can only use an internal netdrive program to sync my files back to home. For now Logseq MD works fine this way, but I’m a bit worried about multi-device sync for Logseq DB version. I have had experiences that the netdrive program caused corruption of a sqlite db file.

Also about the Logseq DB web version, does it opens the graph on the server or a local graph?

1 Like
23

Ah I see. No I have actually played around with NewTags quite a bit myself so I am with you. I thought the were separate from NewTags but if they’re just part of the package so to speak then that makes sense.

25

I do hope they update the UI of both but I suppose that might not be possible practical.

The UI isn’t horrible but I do find it a bit difficult to use purely and simply because the various buttons you need to click are too small.
Also I suppose some things could do with a pop-up menu to make them easier for fat thumbs.

Thank you as always for the update.

1 Like
31

Does anyone know where the DB is stored when you create a DB graph in the web client?

One of the things I love about the current MD version is that I can use demo.logseq.com in my locked down work environment, and have the markdown files stored locally. I can create a new DB using test.logseq.com but it doesn’t ask me where to save it, nor does it ask permission to write files locally.

2 Likes