[Feature] Faceted refinement workflow on top of existing search #24754
Replies: 8 comments 5 replies
|
Immich already supports search like this. All the other bits duplicate existing feature requests. |
|
I agree that the scope and performance of Immich’s search is already one of the strongest parts of the project. From a UX perspective, adding faceted filtering directly within the photo view would unlock a completely new use case: making very large archives genuinely explorable for everyday users. This doesn’t need reinventing. Conceptually, most of the building blocks seem to already exist (fast search + metadata dimensions). It's mainly usability—very similar to what users expect from successful e-commerce experiences: start broad, then narrow down iteratively with clear facets and quick toggles. |
|
Hi @timonrieger, thanks a lot for taking on this topic — much appreciated. While I’m not a developer myself, I can contribute the user and UX perspective based on ~20 years of experience in e-commerce and working with large-scale search systems and search providers. When we talk about filter navigation in Immich, I think it helps to clearly separate two aspects: To avoid turning this into a generic wish list, I’d like to stay very concrete and focus only on what Immich already provides today in the web UI. Current filter dimensions (data side)Many important building blocks already exist:
However, several dimensions behave more like single-choice selectors than true filters:
From a filtering perspective, many of these dimensions would become significantly more powerful with multi-select and basic operators, without introducing new concepts. UX / interaction issuesEqually important as the available filter dimensions is how users interact with them. This is where Immich currently deviates from established faceted-search UX patterns: 1. Activating and deactivating filtersOne of the core rules of faceted navigation is the ability to toggle filters on and off with a single interaction. In Immich:
This makes iterative refinement unintuitive and breaks the “explore → adjust → refine” flow. 2. Missing result countsThere is currently no visible result count, neither:
Without this feedback, users cannot judge:
In faceted systems, showing result counts is essential for orientation and control. 3. Result presentation vs. Photos viewSearch currently always produces a flat result list. As a consequence:
are not available in search results. This makes search feel detached from the primary browsing experience, even though it often replaces it for large archives. SummaryImmich already provides a very powerful and performant search with many relevant dimensions. What’s missing is a true filter / facet navigation layer that:
In other words: most of the hard parts seem to be in place already. The remaining gap is primarily usability and interaction design, not raw capability. |
|
From a UX perspective, filtering and faceted navigation are not “nice to have” features, but established interaction patterns that users expect from any modern content discovery interface. We all use these patterns daily, often without consciously noticing them. Well-known examples illustrate this clearly:
And of course, many of us are very familiar with the comprehensive metadata filtering in Lightroom. These systems all share common principles:
That’s where Immich could really stand out:
|
|
I recently started experimenting with the Immich iOS app on iPad, and I was pleasantly surprised that the filter selection UX feels a bit more user-friendly there. It already looks like an early step toward a simpler “filter / facet” style workflow. However, two very important filters are currently missing on mobile: Tags and Star Rating. In general, I think a consistent UX across web and mobile is in everyone’s interest. The way the apps separate search and filtering might also be a useful direction for future improvements on the web side, especially as Immich increasingly serves larger libraries. For context: @timonrieger — would love your thoughts on whether aligning web/mobile filtering UX is something the project would consider as part of this broader effort. |
|
Hi @bo0tzz, thanks again for taking the time to review and moderate feature requests — I really appreciate the work you put into keeping Immich focused. Since the initial post, the topic has been discussed in much more depth and has produced a fairly concrete, UX-focused breakdown (what works today, where the gaps are, and how this could evolve without reinventing search). In my view, that conversation is now a valuable input for Immich’s roadmap rather than just a broad request. Additionally, a recurring theme is the importance of consistent usability between the web UI and the mobile apps. Aligning the interaction model for filtering/refinement across platforms seems like a key quality aspect as Immich continues to mature. If reopening is acceptable, I’m happy to keep it as a consolidated “umbrella” discussion that links out to the more specific existing feature requests, so it doesn’t duplicate or fragment the issue tracker. Thanks for considering, and for all your work on Immich. |
|
I think most users are afraid of a lot of letters and meanings. I would like to raise this topic in relevant |





Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
I have searched the existing feature requests, both open and closed, to make sure this is not a duplicate request.
The feature
Immich already provides powerful search and a basic set of filters.
This request is not about adding new metadata dimensions. It’s about evolving the existing search/filter UI into a more “faceted refinement” workflow: users can iteratively narrow down large result sets with persistent, easily adjustable filters (explore → refine → adjust), with minimal UI friction.
Motivation
For large libraries, timeline browsing and free-text search alone are often not enough for efficient discovery. Immich’s search is strong, but the current filtering workflow is still fairly basic and makes iterative refinement harder than it needs to be—especially when working with large result lists.
Users need a structured way to narrow down large result sets and explore the library efficiently.
Proposal (UX / workflow on top of existing search)
Build on Immich’s current search and filters and improve the interaction model to support faceted refinement:
MVP goals:
Non-goals (out of scope for this request)
Thanks for all the work on Immich — it is an impressive project.
Platform
All reactions