Probably a good idea to change the name now... However, I'm always surprised people don't search for the name of their project before using it. Also, if common TLDs are already registered, I'd usually avoid using the name as well in case of future success and spoofing the domains.
It is often good for the ecosystem that we have choice of tools for a given task, so I applaud the general effort gone into this additional "Google Photos 'Clone'".
Here the challenge is to actually understand what "Google Photos" is before claiming that this is a Clone.
In order to successfully clone Google Photos, one needs:
1.- A backing storage.
2.- A web application.
3.- A mobile application (and optionally a desktop application).
4.- A data synchronization model to maintain a 2-way sync between backing storage and each locally installed application.
5.- A data model of pictures, timeline, and albums, that does not depend on where the picture is physically stored (for example, I can create an album on my phone offline, and whenever I come online the sync happens and the same album appears on the web application view).
6.- AI features (eg. face detection) sprinkled on top of the backing storage, and available on both web application and locally installed applications.
From a cursory review, this current attempt only satisfies 1 and 2.
7.- A method to proactively decrease quality of all images to save storage, and delete all images if you don't look at them for 2 years or have too many.
Unless Cloudinary is open-source (hadn't heard of it before, a quick browse of their site makes me think it isn't) then calling this a Free/Open Source project is somewhat questionable, since it leverages Cloudinary for a lot of what it does.
My only gripe with Immich is it renames the files to UUIDs, and stores them in its own directory structure. This works ok if you are only backing up images from your phone as file names don't make sense anyway, but if you have existing backups it makes it harder to sync them.
I am planning to use NextCloud for the sync, as I like its YEAR/MONTH/FILE structure, and configuring it as an external library in Immich. I could also use NextCloud Memories, but Immich UI is nicer, and its face/object recognition model is better.
Historically there were integrity issues with the storage template, these should now be resolved however. The reason is disabled is quite simply that most users don't need it and it can cause a lot of IO in some scenarios, for example if your upload and library directories are on different mount points.
External libraries are no longer readonly in the latest release as many users wanted to be able to delete photos in external libraries, but you can mount the Docker volume containing your external library as readonly if the readonly aspect is important to you.
Yeah, saw that note earlier in the week. I have it set to read-only through docker for now, but I like the idea of being able to edit metadata. We'll see.
Unless you explicitly want write, I'd leave it at read-only anyway since they had a (very, very quickly corrected) bug that could result in data loss for external libraries not mounted as read-only this last week.
> but you say it can also use external libraries? so it does not copy those?
Yes, it only renames files automatically sync'd from your phone. My understanding is it creates a separate metadata file for each external file, but doesn't touch the actual files in an external library.
it's not uncommon for Open Source projects to rely on non OSS services. the project itself is Open Source which is where the designation comes from
as far as free, it's absolutely free to spin up. it does require registration (like any other service), but Cloudinary has a generous free tier. it was made pretty clear on the website and repository it relied on that service
just for transparency, i work at Cloudinary, but I stand by the above whether it was Cloudinary or another service
I am surprised by the over-engineered comment. Object auto-detection is kind of an advanced feature and I would expect a lot of extra infra required for it. Immich hides it because they just publish docker containers for you but I imagine it would feel more complex if you were to look at the internals.
> offline ... with object detection search ... All I want is an app which I can point at a folder and then search for "cat" or whatever and have it show all the pictures with cats
From yesterday on HN:
Desktop Docs is a Mac app that lets you search all your photos and videos in seconds with AI.
Once you find the file you're looking for you can resize it, export it to Adobe Premiere Pro, or drag and drop it into another app.
I built Desktop Docs because I keep tons of media files on my computer and I can never remember where I save stuff (lots of screenshots, memes, and downloads). The Apple Photos app also only supports photos in your iCloud.
Desktop Docs supports adding folders or individual files to an AI Library where you can search by the contents of your files, not just file titles.
You can search by objects ("cardboard box"), actions ("man smiling", "car driving"), by emotion ("surprised woman", "sad cowboy"), or the text in the frame (great for screenshots or memes).
It's also 100% private. Make any media searchable without it ever leaving your computer.
Check out https://1lilac.com. Still in development. Pricing is temporary. Keen to hear feedback. This likely falls into the "vastly over-engineered" category. Also, it is only online at the moment. Disclaimer: I am the developer.
Probably a good idea to change the name now... However, I'm always surprised people don't search for the name of their project before using it. Also, if common TLDs are already registered, I'd usually avoid using the name as well in case of future success and spoofing the domains.