Help! I need a PKM solution that will allow me to build a database of images of works of art that includes metadata on the work itself as well as notes and possibly attachments. My goal is to be able to produce a gallery view of the images based on search terms like artist, genre, dates, location, what it’s an example of, etc.
I currently use Obsidian, but I’m more than open to using another app for this particular purpose. I don’t mind paying if need be, and I don’t need to collaborate with anyone. Privacy in this particular case is not a concern. I will not be using the app as my only repository for the information I put there: copies of the images and the relevant data will be safely tucked away on a physical drive. If the app blows up, I won’t have lost any data, just the labor of setting it up.
Thoughts? Is this something I can make Obsidian do via Bases? I’ve tried doing this in Lightroom, but it’s not ideal since I’m limited to tags and exif data for metadata and filtering. How about Notion or Anytype? Is there something else I should be looking at?
This is the correct answer (and from my research the ONLY correct answer.)
It’s an absolute godsend for organizing your visual library. One thing I will add - they don’t have a phone/iPad app, but there is an app from another dev that works with Eagle’s database format called Hive. Fair pricing, one time purchase option for both.
Below is a screenshot. I pasted a few random photos into the inbox. As you can on the right, you add all kinds of metadata and create smart folders from that information.
I did give Tropy a try a few years ago. It is almost perfect for what I want to do, but for one thing: metadata fields are limited to those available in controlled vocabularies such as Dublin Core or BIBFRAME. No problem with that, but unfortunately it no longer appears to possible to download and import the available vocabularies into Tropy. (I keep trying, and I keep getting error messages.) What’s available without adding additional vocabularies is pretty limited, alas.
Thanks for the suggestions—I confess I didn’t have something like Tapforms on my radar. I note that they are all subscription apps now. I’m not averse to subscriptions by any means, but I’m not sure these apps will add enough value over what I can already do with Devonthink (or the free versions of Notion or Anytime) to justify using them for what will be a relatively small project.
I’m on the cheapest plan of mymind (about 4 USD a month) for saving and organizing images. It’s good for grouping images, writing notes about them, and there’s AI tagging. All files and notes can be exported. It might be too much for something I do for fun, but it’s frictionless enough. I don’t mind paying for the subscription.
Yes you certainly can do it in Obsidian, but it will take a bit more work. I’d start by looking at how folks are implementing book catalog data with cover pictures and notes into Obsidian as a starter.
For me I’d spend time initially on deciding the fields you want in the properties and then move from there into using bases to slice and dice the data.
However, I’d actually suggest you look at art museum collection management systems instead. One of the good open source ones is Collective Access but there are other choices. Omeka is one and I think there are a couple of others. They are designed specifically to collect and store the types of metadata you are looking for and provide robust ways to search and edit the data.
Quite a few of the images I need to manage are already embedded in Obsidian notes with data-rich YAML headers and links to supporting notes and documents. My ideal solution would be porting those notes into an app that uses the YAML I’ve already crafted and retains the links to the notes and documents.
I’ll check out the museum collection management systems to see if they can accommodate linked notes and documents, which, the more I’ve flailed away at this, seems like a key part of the puzzle.
I think MyMind is beautiful. I gave it a test run about a year ago—what it’s missing that I need in this particular case is user-defined properties. Tags can’t do the work that properties do unless you can create an orderly system of nested tags.
If much of the information is already in Obsidian I’d stay there and spend the time and effort learning how to integrate what you have using Bases for your outputs. I know I’ve seen example Bases documents with gallery views based on criteria in some vaults so I know it can be done. It sounds like you already have a good percentage of the work done if you have a decent YAML set-up.
Apparently I could also import Obsidian notes into Anytype, which will then convert YAML headers into properties and retain tags and links. I’d prefer to stay in Obsidian, which is my main note repository, but if I can’t make it work, I’ll give Anytype a try since I won’t lose the work I’ve already done.
A little late to the party but if Obsidian doesn’t work out for you, may I suggest TiddlyWiki? It’s a venerable, but seemingly little heard of, player in the PKM arena. Having been around since September 2004 it is well rounded and stable.
Its basic unit of information is called a “tiddler”. Tiddlers can contain text, images, PDFs, anything that can be rendered by a web browser. They can be linked, with automatic back linking, and can have open ended, user defined metadata attached (ie you can define the attribute as well as the attribute’s value).
I think, however, the most interesting aspect for you may be TiddkyWiki’s tagging system. Tags are themselves tiddlers, meaning they can be described in detail and given a parent. That allows tag hierarchies to be built, which in turn allows proper set based queries. If I have a photo of my cat Tiddles that I tag as “Tiddles” and I make the parent of that tag another tag “Cat” and the parent of that tag “Animal”, (an appropriately formed) query for any of the three tags “Tiddles”, “Cat” or “Animal” will return the picture.
There isn’t a gallery view by default, but there are extensions that can be installed.
Disadvantages? It’s a browser based and for the type of work you’re looking at you’d need to install the version that runs under node.js if you want everything on your system/ lan. There is a quick and easy version that is literally just a standalone file, but start throwing in a catalogue of hi-def images and the file will get too big to be manageable.
Also, although it understands Markdown, the preferred language is a more powerful and older one called WikiText.
If Obsidian doesn’t do the trick, I will check it out. (I’ll probably check it out anyway because you’ve piqued my interest … I’ve loved poking at these kinds of systems since the glory days of Lotus Notes. )
)