Former Immich dev here:

Even in the early days, we got hundreds of different feature requests that “can be done easily”. Even this example is a lot harder than you might think.

• Since all clients work differently, this probably has to be done server-side, requires a new endpoint, caching logic, client support.
• Immich supports a ton of file formats, the solution has to work with all of them.
• We must not lose meta data, quality or alter the image in some other way, Immich is a backup tool, file integrity is a top priority.

If you have an idea, how to tackle all of this, contact the dev team on Discord, they’re really helpful and will surely guide you through the process of implementing.

The problem here is, that the server does a lot of image processing to generate various thumbnail sizes. This is the main reason images in the gallery load so fast when scrolling. But to archive this, the server needs to read the image which would not work with e2e.

I was one of the first Immich core developers but had to leave a year ago due to a change in my job situation.

Just wanted to say that the team (especially Alex, the founder) are wonderful people and very dedicated to the project. When they don’t implement the features everyone wants, there is usually a very good reason, like security or architectural concerns. The philosophy behind the project is also to do things slowly but right.

I believe they continue focussing on robusness and bugfixing. A topic that we discussed a lot was e2e encryption. We really wanted to implement that, but it was very hard and back then would have compromised performance too much. I still think that e2e encryption will come at some point.