Repository navigation
Exploring Immich Mobile for iOS #33515
Replies: 72 comments 57 replies
|
1000% the way to go. Hopefully you can work with NAS vendors to make app native as well. |
|
I think this would look amazing! Having an effective design makes the app very integrated with the iOS Ecosystem. |
|
I love this design a lot already, it looks beautiful and seems to run smooth like butter. |
|
Collapsable/expandable tags tree please 🙏 |
|
I don't really care about HDR. I smaller app size would be nice to have too. |
|
That is great news! Looking forward to this. Immich is already great, even with the flutter app. However a few things that would be nice:
That's a quick list of what I can think off just now. |
|
I see the benefits and how it might be the right approach, a little sad because departing from a single codebase might result in feature disparity with Android. There are some apps out there that are great on iOS and feel abandoned (or just ugly) on Android. |
|
I am hoping that with a native iOS app, Immich could take advantage of new metadata exposed by the system that previously wasn’t available with the Flutter API shims. And maybe fix some of the time zone issues, especially with screenshots 😄 I think this is a good direction. Always a big fan of tighter integration with platform features and idioms. |
|
Is this enable new siri Ai? |
|
Amazing work so far Adam with the proof of concept. I've been using Immich since the beta days and I’m excited to test out the new iOS app when becomes available in TestFlight. Cheers ! |
|
I'm very curious about the "Zoom"-grid demo here, and what work might have been put into getting that working behind the scenes? I've long been following @Sparklysparkspark's work in #28696 & #28641 in hopes of seeing an ungrouped overview grid for Immich, as it's been one of the few holdout features I've been hoping to see added for a long time. If work on getting a grid-view is already happening behind closed doors, I'd love to see it surface so that it could be explored for other platforms as well! (As stated before, I'd also be happy to sponsor any work on this if so desired) |
|
First off, you guys are doing an amazing job, and a native app would be amazing. |
|
Looks fantastic and I’d be very keen to help with the TestFlight however I can. Would this allow integration with some more core iOS features, such as tap & holding an image to cut a person/object out? Or even some of the Apple Intelligence (or Siri AI?) features like photo indexing, or even the object removal/cleanup tools? |
|
Exciting! I would love to be involved with the Android app if and when it gets started — I’ve been building primarily Compose UI and Android app architecture for the past several years across two companies. Not sure if you have any Android engineers on your team, but something I’d be happy to volunteer my free time on |
|
As a long-time Android user, I hope Android isn’t sidelined in favor of a wonderful iOS app. Unfortunately, there are often extreme differences between iOS apps and Android apps—even from major manufacturers. I’d like to see full support for tags in the mobile app. I consider this an underrated feature. I use tags to keep my Google TV screensaver refreshed with new content. |
|
This might be the happiest moment for me since I started working on Remmich, a Readonly-Immich client with native iOS support. Canot wait to try out your POC on native iOS client. At the mean time, I wonot stop working on mine tho, it's nearly finished, just one and half milestones left, even tho I am pushing it after work, during weekends, but I enjoy the process of turning ideas into a "product." 💜 |
|
Just publish it already !!!! |
|
You should focus on android really, it's had for more users. |
|
This would be absolutely amazing! also if you could add it so that the server update is available links to the App Store instead of GitHub too. thanks! |
|
Does this also mean we might finally get an Apple TV app? |
|
Opportunities:
Risks:
Questions:
Advice:
|
|
Id would like to see on the iPad side the Navbar more expanded (like on the native app) to see the albums and people (and to pin and sort these categories) even on the collapsed form. |
|
I would love to try immich when the native iOS app is out! This is great news. |
|
Love this as an idea! I am curious to see if it might not be easier to write the updated Android app in Kotlin Multiplatform, and then SwiftUI for the iOS UI, and Jetpack Compose for the Android UI. Then it's one codebase to manage. |
|
I have iPhone 13 and I'm excited about this. Keep it going. 🚀🚀🚀 |
|
I would like to delete a memory so that it's not shown in the memories widget, similar to iphone photos. Currently we can only click a memory and view its exact location in the album instead of a collated stack in iphone and the recommended way to remove the memory is to archive the photos individually but I feel like archiving is a way of classifying images |
|
Coool, a suggestion from me would be on the ios app to be able to do the same as on the android app regarding saving if possible (by chosing to upload only when on wifi + charging) |
|
@agg23 with the new IOS native app will this issue be addressed #5818. Will working in swiftUI and UIKit make pulling original photos with full metadata easier? with Immich's main goal being to back up your photos i feel that this is quite important especially if a lot of work is going into making a new UI for apple. IMO the UI means nothing if the backing up doesn't work as intended. |



Uh oh!
There was an error while loading. Please reload this page.
Greetings, I am Adam (agg23), the newest addition to the Immich team, where I work on mobile. The Immich team has been growing (two new engineers in the past ~6 months), allowing us to allocate more resources into making the Immich mobile experience even better. We have a number of projects being explored in parallel at the moment, and today I would like to show off our proof of concept of a native iOS app.
Why Change?
Immich mobile is built using Flutter, a cross-platform framework for building applications. Flutter has a number of nice properties that have allowed us to work very quickly, developing the Immich app to what it is today with only one or two core contributors. However Flutter has some downsides, such as no native HDR support, non-native UI/interactions, and general performance issues. These are things we could work on solving through Flutter, but we wanted to investigate more of our possible options.
On top of these problems, Flutter exists as "odd platform out" on iOS. It works right out of the box, but Flutter does all of its own drawing and doesn't react well to platform changes, as we recently saw with the release of iPhone Duo. When Duo was announced, we immediately noted problems in Immich with safe areas and unconventional control placement. These are at least partially our fault, and partially Apple's fault, but the fact that we have to revisit scenarios like this weighs on us and any solution we may build. Along similar lines, we cannot support HDR without patching a native view into Flutter, as Flutter doesn't support the higher color space, which hurts our image viewing experience significantly.
A native version of Immich solves these problems for "free" and lets us focus on building new, interesting features rather than reacting to a third-party platform.
Introducing Immich Native for iOS
After considering our options for solving problems like HDR support and iOS UX particulars, I began prototyping how we might reimagine Immich if it was specifically built for Apple platforms. We are thinking a lot about our visual identity and the desired user flow (for example, how can we improve on proprietary platform’s viewing experiences?). Due to the ease of human building, most of the UI is built in SwiftUI, with the critical sections (read the timeline and animations) written in UIKit. Automated tooling seems to prefer UIKit, however, so we may reconsider this as time goes on. The native iOS UI is built around a compiled core library (Rust) containing all of our logic, state, server interaction, etc., which is largely a new construction.
We’re still early on, so what you see in this post is a proof of concept only; do not expect it to represent our final intention, design, or performance targets for the project.
What's New?
Revisiting Immich from the perspective of the Apple design language gives us new opportunities to think about what the Immich design philosophy should be, especially with our different-than-corporate goals of being privacy-centric and self-hosted. We value consistency, so we don't want to change things just to change them, but we also don't want to artificially limit ourselves if there's a better alternative. For example, this has led into conversations about where and how to surface memories in the app. Memories are one of those key emotional parts of Immich that we want to drive people towards, but at the same time we don't want to get in our users' way.
The app is very early, but we're already seeing some nice benefits:
Additionally, we have some other core goals that we are working towards:
We haven’t done much performance optimization, but this is what scrolling around on an uncached 200k picture library looks like:
Bottom.Perf.Animation.mp4
Bottom.Zoom.Demo.mp4
This may not make it into the final product, but I'm particularly happy with this new animation:
Bottom.Animation.Demo.mp4
What's Next?
We still don't have a huge team, so this is going to be a process that takes some time. We plan to begin offering this app via TestFlight as soon as we have a minimum feature set, and it will remain under a different bundle ID for the duration of its testing period (thus allowing you to use normal Immich and this new app side by side). It will be merged into the main monorepo relatively soon alongside the current mobile apps. Of course as soon as it's in the repo, you will be able to build and play with it yourself regardless of TestFlight.
One of our big goals is to make backup as solid as possible. It has been a journey for Immich to handle the ever changing quirks of iOS backgrounding and PhotoKit access. I want to take our years of experience and funnel them into building a testing solution that gives us more confidence over the sync and backup mechanisms. As we get to revisit these segments of the codebase, the opportunity arises to fix long standing bugs (and introduce new ones), so hopefully we see some big improvements here quickly.
I have a lot of work to do to polish Immich's mobile UX. I want to make the experience as smooth as possible for the user; making it very clean and intuitive. iPhone Duo has some very fun and interesting problems to design for that I've been experimenting with in my free time, and I'm excited to see how we can make Immich the best photo viewer for Duo (amusingly, Apple Photos regressed with 27.2 on the Duo, so maybe beating them isn't so hard).
This is a fun Duo experience I’ve been playing with and haven’t seen anywhere else:
Bottom.Duo.SideBySide.mov
What About Android?
Despite all of this talk about iOS, we want Immich Flutter (Android) to get the same share of attention. The native iOS app is built against a shared FFI core, which we plan on eventually sharing with Immich Flutter. This is work we were already planning on regardless of this iOS project, but it fits together with this PoC nicely. Once this work is completed, all of the new performance and stability goodness we’re building for iOS will come to Android for free.
Android will be (eventually) getting all of the rewritten and tested logic as we begin to share more and more code. We don't have the resources to overhaul the app for two platforms at once, and iOS makes it very obvious when something is not done the "Apple" way, so we’re starting in that direction. We may pursue native Android UI based on the success of this venture, but we also may take the learnings from this project and use them to improve our existing Flutter codebase.
Tell Us What You Think
We are very interested in your feedback. We do not take rebuilding the mobile app lightly, but after some thought and experimentation we think this is the best path forward.
Any comments you provide will factor into our decision making as the project continues. Our community is our largest asset and we strive to build the best experience for everyone.
All reactions