Is Devonthink as janky as Devonsphere? CLI interface? Lock-in. Alternatives

Problems with Devonthink:

  1. Devonsphere Express often lags when I use it. I notice that it looks like it’s programmed in a way where the whole thing is single thread when trying to update because you can’t use the app while it’s trying to update. It was unreliable when I used it back in 2015, and it doesn’t seem to have changed much, giving me a spinning logo when I try to hover over its icon. Is Devonthink like this? Is there a more reliable CLI interface by any chance? Is there any plans for a total rewrite using modern AI tech we didn’t have before?

  2. I think it can’t export the metadata it’s produced to work with other apps? Obsidian shares its data, as does HookMark, but not DevonThink, nor Eagle Filer. So AFAIK, there is no alternative?

  3. Not cross platform, so locks me in to Apple. This is a bummer. I also won’t ever get an iPhone because the walled garden would have affected me in the past.

Thus, I wish there was an alternative!

At the moment, I’m just getting by naming all my files as verbose as possible, and finding them with Raycast. Then trying to get myself to store them when I’ve accessed them more than 3x. I use ObsidianMD for notes, Pages for Presentations and SkimPDF for PDF presentations, with occasional use of Hookmark.

If “lock-in with Apple” is a bummer, then why use Pages, Hookmark or Raycast? What exactly is your point, if it’s not only venting your feelings?

If you mean the internal index – most probably not. That’s their business secret, so why would they want you to export it. And what would you want to do with it? If you mean metadata that you added like tags etc., a closer look might help.

Alternative to what exactly? To not exporting some unspecified kind of metadata?

But that’s great – you’re getting by with the tools of your choice. Congrats! I still don’t quite understand what exactly your post is about if you’re happy with what you have.


Well yes, the moan is that there’s no alternative. I can put up with a few of these problems, but there’s just too many.

Look at the way ObsidianMD works. The vault is just plain files. This means you get interoperability. If Obsidian can’t do it, I can process those files using sed, grep; all kinds of things. That’s great.

The nice thing about Devonthink seems to be that it’s mainly a search layer, but it looks one way. Or am I wrong? Is there some interoperability I’m missing, like working with Alfred or Raycast? Can I use it via a 3rd party app like this and avoid the UI?

Also, it’s not a rhetorical question: is Devonthink a reliable app that never freezes? As I say, I bought Devonsphere, but it didn’t match searches as well as Raycast. That’s a shame because unlike Raycast, Devonthink can help organise and find connections, which is what I’m looking for.

My system isn’t a disaster. As I say, I get by from using filename search. However, there are times when it becomes a problem.
One of those times is when I need to work with other people who aren’t privvy to the system. That’s where Devonthink could help. Or Devonsphere, if it was designed for that.

As I said: If it works for you, good. If you’re looking for something else, you might want to ask questions. If you just want to moan about supposed shortcomings of DT, you’re perhaps not at the right place.

As to your questions: The forum has a search function that will answer some or most of them.

Whoa. Is there any reliable app that never freezes? If you find that, stick with it.

This forum is great at answering concrete questions (after you’ve used its search function) and providing help with concrete problems. It’s not so good at dealing with overly broad and general allusions or problem descriptions in the sense of “I want to find something somewhere, is DT the right tool for that”.

  • No, there is no CLI for DEVONsphere. And we don’t publicly comment on development plans of timeframes.
  • No, DEVONsphere doesn’t export its metadata or hook into other applications.
  • We are Apple developers. Despite any complaints we have about decreasing quality control under Tim Cook, it’s still the best of breed for us.
  • What have you added to the Categories and Locations in DEVONsphere? Hopefully not your home directory or entire hard drives, especially networked ones. That is not what the app is designed for.

No networked drives.
If I try to index everything: 873693 files. 518gb. I guess this is too much?

Alternatively, would this be possible? Just indexing this:

Old Docs, folder system: 30941 files. 66gb.
New system, tagged files: 14590 files. 180gb.

Devonthink has some cool features with Hookmark that are really useful and appreciated:

Your second suggestion for indexing sounds like a better idea. But it would also help to know what you’re trying to accomplish.

Also, were you trying to index those 873693 files with DEVONsphere?

What is it that you want to achieve?

Over the decades the only true constants that will guarantee you future proofing and system interoperability is plain text files and their file names. The moment you start thinking about adding metadata outside of these 2 things, you start getting “locked in”, regardless the operating system or app you are in. Even xattrs are not future proof as filesystems evolve over time, ditto to many plugins in Obsidian.

One other constant is perhaps the directory or file structure, so a nested structure will give you another “dimension” of portable metadata.

If you use graphics, sound or videos in your files, that already is brittle in the context of interop and future proofing.

If you really want to make sure that files you create now are usable 100 years later on some yet-to-be-invented system then I would suggest using vim (or other such plain text editors) and work with plain text files with none-to-minimal markups eschewing any form of graphics (except ASCII art), sound, video etc. To make presentations, do it the way we used to when all we had were text terminals – use plain text files, and count lines to separate out “slides”.

I would also not shirk from using XML or JSON (or CSVs for tables) since they are nothing more than work-arounds to increase the dimensions of flat text files with nothing more than text, and you or your successors should be able to create scripts to manipulate them where ever they end up.

This is not some slight on your ideas, it is something that I have been thinking about as well having used everything from VAX to unix to NeXT to BeOS (remember that?) to Windows to OSX etc, etc. and numerous filesystems. It really comes down to what you want to achieve, which will in turn inform you on what compromises you need to manage.

I am a heavy user of DT3 but I know that if I am going to switch away from the Apple ecosystem one day, I will easily be able to extract my data but I’ll have to discard some metadata that Devonthink3 provides for convenience such as ratings, colour labels and Devonthink3 tags. It’s why I make minimal use of them except when I need to do something efficiently now and those tools allow me to quickly, with minimal effort, manipulate my files.

I also place important metadata in the filename. All files have a date or two in its name since even creation- or modified datestamps are not sacrosanct, and sometimes context (e.g. project name at the end). I make full use of the 255 bytes.

My philosophy is to make use of the best tools available now but keep the underlying data that I want to preserve as simple as possible, knowing that a majority of the data I receive or generate does not need to be preserved (since I’m not someone future generations will want to study!).


1 Like

I would suggest using vim (or other such plain text editors)

I have been using CotEditor as my secondary editor for formats like XML. Just sharing a suggestion for a nice app :slight_smile:

My goal:
I already have my new system of naming files verbosely and simple search. However, there are times when it doesn’t work. That’s when I’m looking for help from Devonsphere.

For example, let’s say I’ve made a typo when naming the file, so when I search for it, I can’t find the file.
I’d like to be able to use Devonsphere to help me find the file by searching for similar files and having the file with the filename typo come up in the search results.

That’s all I’m looking for, but I’m not sure how much Devonsphere can handle. I’ve updated settings to just index the 2 filing systems and not the entire drive and it’s handling things better when not updating.
( Old Docs, folder system: 30941 files. 66gb.
New system, tagged files: 14590 files. 180gb. )
I think once it’s settled in from indexing it’s more reliable. But I need a guideline for keeping the data below some sort of limit, both number of files and total size

Other disk-level search tools exist, fwiw.

1 Like

Thanks! Some great advice there.

This led me to:
Linking to obsidian: GitHub - ryanjamurphy/DEVONlink-obsidian: Open notes indexed in DEVONthink in, well, DEVONthink but I don’t see it as part of the solution. Sharing that, but shelving for now.
Command line tool: GitHub - mhucka/ask-devonthink: Ask DEVONthink for specific metadata field values of items selected in the front window
Alfred: GitHub - zeitlings/alfred-devonthink: DEVONthink 3 Portal
Raycast: Raycast Store: DEVONthink

Hookmark: I use it. Ties to MacOS though.

EagleFiler. I get confused with this and another app called Eagle for visual / photo organising…? Stores files in open format; good. Could paid well with Devonsphere indexing anything I missed. MacOS only.

Perhaps a combo of Obsidian, EagleFiler, and Devonsphere could be the way to go.

This is a spam bot, probably AI based (the joys of progress).

1 Like

“This” is “johsmith3”? Then best to delete all this as it’s all rubbish talk that is out of talk, frankly.

We did in the meantime.

1 Like