Hey everyone! 👋
I'm a blind Immich user and the app works great for me, especially after the last mobile update that now reads dates for photos in the gallery viewer. I wanted to suggest a couple accessibility improvements for Android/web that would be game-changers for blind users.
🕐 Better date/time info Right now screen readers only announce the date, but I take tons of photos throughout the day. Would be amazing to have "Image taken in [city], [country] at [date and time]" - just hours/minutes would be perfect!
📝 Image descriptions in ALT text I have a workflow where I use Gemini Flash to auto-describe my photos, plus Immich's face detection = perfect memory system for me. But currently I have to open each photo → tap info → read description.
Would love ALT text like "Image taken on [date/time]. [description]" so I can browse the gallery and instantly know what each photo shows.
I think this might be partially working already (Android seems to cache descriptions after viewing), but you have to open each photo first.
Planning to try implementing this myself since I'm learning Flutter, but wanted to check if this sounds doable without major refactoring first! 🤔
Thanks for reading! Let me know what you think.
#Immich accessibility improvements
plucky phoenix
gusty sigil
Hey! Thanks for sharing and I really appreciate the input. We're aware that accessibility currently is far from perfect, which sucks. For the image descriptions we have in fact discussed possible routes for a local ML model that summarizes pictures so that we can use that summary for alt texts. Unfortunately that's not coming anytime soon soon but we'd like to have that eventually
Generally getting input on this from visually impaired people is very helpful, since nobody in the team actually requires accessibility features. Thanks!
plucky phoenix
You're welcome! I'm glad to read you're interested in this. It's also a bit hard for us to have photos, video software adding accessibility features :). Right now we have lots of things for generating descriptions: there are apps (such as Be my eyes, seeing AI) that can describe a photo and allow you, as a blind user, to copy that description. There's also Gemini, which is free or cheap and generally can describe video too. So I think if you can make the app to add the description into the ALT text we could add descriptions by ourselves with this kind of softwares. I've been thinking in releasing a simple app that could be used just for this, using the Immich and Gemini API'S.
ornate yew
Hi - thanks! Very valuable input! We would really value your insights here. One question I had: how do you feel about website that have an 'accessibility mode' (floating button). Would you rather have it just "always accessible" or would it be ok to add additional acessibility features by explicitely enabling them?
Example: https://apnews.com/
Also, for things like alt text - would you rather have complete sentences like: Image taken on [date/time]. [description] or more like bullet points: Time: [date/time], Description: [desc]
plucky phoenix
Generally speaking, I think if these changes require more API calls (example if you need to retrieve exif data for accessibility, and you don't do that currently) or generally a different web layout and stuff it would be OK to make them under a new mode. That mode however would be cool if can be remembered. Also there could be a setting in the user profile that enables accessibility. I think some apps implement this already for certain features. I wouldn't name this as "accessibility mode" or something because the app is already accessible, but this might be like accessibility improvements or something like that.