That’s a fair bit of skepticism. In this case it’s justified. It’s used to auto-tag photos, either cats, dogs, people, buildings; or if you manually tag people in some photos it will go through and auto-tag your whole library. It’s all done locally on the server, no calls out to external services, so it’s private. And it works pretty well.
So, yeah. In this case, the ML is useful and justified.
It needs a server. You need a server of some sort of you’re synching your photos off the phone. If you’re only looking for a different gallery phone app, there are several good ones (and a few interesting, novel approaches) in f-droid.
They do have an official app. I used to use the fork, but I changed my workflow and didn’t need the patch anymore and went back to the upstream version of the server and client.
Huh; thanks, I must have grabbed the wrong one when I switched phones. It’s been working fine, but then, I don’t use it much. Good to know I need to switch, though - thanks again!
For phone photos, I use a FOSS gallery app. For photo syncing, I use PhotoBackup, plus a script on the server that watches the upload directory and calls the PhotoPrism “Import” function via the REST API.
Thanks, but I wonder for how long this will continue to work, considering that the latest PhotoBackup build is from November of 2016 and it targets Android 4.1 (newer Android versions will probably require a higher version target). In the long term, I think that migrating to Immich or Ente is the better option. Both have official, well-maintained FOSS apps, that are available on F-Droid, and the backend services can be completely self-hosted.
Android could break it, but I don’t subscribe to the philosophy that an essentially bug-free, functioning piece of software needs constant updates to stay useful. If Android doesn’t break it with API shenanigans, I’m perfectly happy to continue using it.
If Android does break it, all it is is a file syncing service. SyncThing could slot into the workflow with minor effort.
Modular systems rock. They have many advantages over all-in-one, monolithic systems; that they’re harder for users to set up and use is almost the only way they’re inferior.
PhotoPrism. I’m not convinced it’s the best, but it works for me and I’ve zero motivation to switch.
“AI Powered” :/
Exactly what I’m trying to avoid.
But I’ll consider it if it’s opt-in only and it doesn’t nag me.
That’s a fair bit of skepticism. In this case it’s justified. It’s used to auto-tag photos, either cats, dogs, people, buildings; or if you manually tag people in some photos it will go through and auto-tag your whole library. It’s all done locally on the server, no calls out to external services, so it’s private. And it works pretty well.
So, yeah. In this case, the ML is useful and justified.
Ok, thanks. So it needs a server, it seems? Can it run as a standalone phone app?
It needs a server. You need a server of some sort of you’re synching your photos off the phone. If you’re only looking for a different gallery phone app, there are several good ones (and a few interesting, novel approaches) in f-droid.
Cool. Thanks!
Which app do you use with it? This one? As far as I’m aware, PhotoPrism doesn’t have an official Android app.
They do have an official app. I used to use the fork, but I changed my workflow and didn’t need the patch anymore and went back to the upstream version of the server and client.
The app you linked to is unmaintained and archived on GitHub, and the Readme says:
https://github.com/thielepaul/photoprism-mobile
Huh; thanks, I must have grabbed the wrong one when I switched phones. It’s been working fine, but then, I don’t use it much. Good to know I need to switch, though - thanks again!
For phone photos, I use a FOSS gallery app. For photo syncing, I use PhotoBackup, plus a script on the server that watches the upload directory and calls the PhotoPrism “Import” function via the REST API.
Anyhoo, thanks for catching that.
Thanks, but I wonder for how long this will continue to work, considering that the latest PhotoBackup build is from November of 2016 and it targets Android 4.1 (newer Android versions will probably require a higher version target). In the long term, I think that migrating to Immich or Ente is the better option. Both have official, well-maintained FOSS apps, that are available on F-Droid, and the backend services can be completely self-hosted.
Android could break it, but I don’t subscribe to the philosophy that an essentially bug-free, functioning piece of software needs constant updates to stay useful. If Android doesn’t break it with API shenanigans, I’m perfectly happy to continue using it.
If Android does break it, all it is is a file syncing service. SyncThing could slot into the workflow with minor effort.
Modular systems rock. They have many advantages over all-in-one, monolithic systems; that they’re harder for users to set up and use is almost the only way they’re inferior.