KDE Discover Redesign: What Actually Needs to Change (and What Already Has)
KDE Discover has a discoverability problem
That is not wordplay. Open the app without a specific name in mind and there is almost nothing to pull you in. Discover is KDE Plasma’s built-in software center, handling Flatpak, Snap, and standard distribution packages alongside Plasma add-ons. It has improved noticeably over the last couple of release cycles. But the home screen still sends most users straight to the search box.
What Discover actually gets right
The Plasma add-ons integration is genuinely unique. You can browse and install desktop widgets, color schemes, Plasma themes, and app plugins all in one interface. No other major desktop software center does that. GNOME Software does not handle GNOME Shell extensions through the same unified flow.
Multi-backend support is real and useful. Discover pulls from your distribution’s repositories, Flathub, and Snap simultaneously. You can pick which source each app comes from, and the repository management screen is simple enough that casual users can actually navigate it.
Recent releases have moved things forward. Plasma 6.7 in April 2026 brought more prominent install buttons and cleaner application page headers. Icon loading was made non-blocking in an earlier update, so the app launches faster and no longer stutters when scrolling through long lists. Small changes, but they add up.
The pain points that keep surfacing
A home screen that shows nothing to browse
This is the biggest one. Without a specific app name already in your head, there is nothing on the home screen to spark interest. Categories are there, but they lead to the same tall list view. No featured apps, no editorial picks, no new-and-notable section.
App stores have solved this for years. GNOME Software has rotating featured banners. The App Store and Google Play fill the home screen with curated sections that surface things worth looking at. For an application called Discover, making browsing feel like actual browsing should be the top priority.
A KDE Phabricator task from 2021 (T12668) proposed exactly this direction: a prominent banner, grid-based card layout, featured categories with screenshots. The mockups were well-received and influenced changes that landed in KDE 5.23. The appetite for going further is still clearly there in the community.
No user ratings in the interface
GNOME Software shows star ratings directly in the app list. Discover shows nothing. When choosing between two utilities that claim to do the same thing, a community score is a useful signal.
Flathub now exposes ratings through its API and shows them on its own website. Pulling that data into Discover for Flatpak results would need no separate rating backend. It is a practical shortcut and one users would feel immediately.
List items that take up too much room
Each entry in the default list view is tall. On a normal display you can see only a handful of apps at once. Scanning a category gets slow fast. A compact list mode or a proper grid view, user-selectable, would help a lot. The KDE Human Interface Guidelines already push for appropriate information density; applying that more consistently here seems overdue.
Concrete proposals worth raising
Build a real home screen
Flathub maintains curation data already. Discover could pull from it to populate a featured section without KDE needing to staff an editorial team. Even a simple trending-apps block would give the home screen something to show. Right now it is mostly empty space that tells new users nothing.
Pull Flathub ratings into app pages
Displaying Flathub community ratings in Discover’s Flatpak results is a narrow, scoped change with real impact. GNOME Software already does it. The data exists. This kind of specific proposal tends to get picked up by developers because the scope is clear.
Keep the install button prominent
Plasma 6.7 made a good start. Continuing that work by keeping screenshots in a horizontal strip instead of a tall stack would make app pages feel tighter. The install button should never require scrolling to reach.
Keep Plasma add-ons front and center
Any redesign that buries add-ons browsing under a generic app store layout would erase what makes Discover different. That integration is the strongest argument for using Discover over just opening a browser and going straight to Flathub. It should stay visible.
Getting feedback to actually land
The KDE design team responds better to mockups than to paragraphs. A rough screenshot with annotations, even hand-drawn, beats a wall of text. Phabricator and KDE Discuss are both good venues. Specific, visual proposals move faster than general impressions about what feels modern.