Graphisoft · Archicad 28 · In-app store
Archicad Add-on Store
One card system for every add-on, in any window size.
What
A store inside Archicad's home screen where users could find and download the add-ons they're entitled to, from one catalog.
My part
The card system, the tag logic, the detail page and the layout rules that keep the store working in Archicad's resizable in-app browser.
Result
Designed and specified for developers. The project was stopped when another team started a very similar product in parallel, and neither version shipped.

01 — context
One place for every add-on
Archicad users get add-ons from Graphisoft, partners and third-party developers. There was no single place to find them, see which ones you're entitled to and download them. The Add-on Store was meant to be that place, built into Archicad's home screen.
We were two designers on the project, and we split the work. I designed the parts every add-on passes through: the cards, the tags on them, the detail page and the layout rules for every window size.
02 — the problem
Different content, any window size
The store had to show very different add-ons in a consistent way, inside a window users resize freely. A card that looks good at one size with one kind of content breaks at the edges: long names, many tags, a narrow window.
Inside Archicad
The store runs in Archicad's built-in browser. The window is resizable, and modal dialogs are technically hard to open there.
Part of the home screen
It had to fit next to the home screen's own left-side menu.
Limited frontend capacity
Frontend development was the biggest risk on the project, so every design had to be simple to build.
Scope for Archicad 28
Free Graphisoft and partner add-ons only. No purchases, no browsing from a web browser.
03 — key decisions
Build on Archicad, design for the edges
A catalog, not one-click install
Our first plan was one-click install: add-ons that download and work right away. That was a huge amount of development work, made complicated by the limits of a legacy system and by who is entitled to what. About halfway through the project, we found that on every forum, users only asked for one catalog where they can see all add-ons, instead of collecting them from many different websites. Installing them and putting them in the right folders by hand wasn't a problem for them. So we dropped one-click install and went on with the catalog.
Start from what Archicad already has
A senior designer had already designed the cards for Project Hub. Before drawing anything new, I studied that system: its sizes, its grid and how it reacts when the window is resized. I kept the grid, the resize rules and the action button, so the store reflows like the rest of Archicad and the main action feels like any other action in Archicad.

A Project Hub card next to the Add-on Store card. Same grid and action button, smaller card. The cards still show the early Install button; the final design used Download.
One card family, not one card
One card wasn't enough for a window that goes from a small panel next to the model to a full wide screen. I designed a family with the same structure: grid cards with one-line and two-line names, a smaller card for narrow windows, and a horizontal row for the list view.
Rules for every width
I defined how many columns appear at which width, and gave every variant a spec: a maximum of three lines of description, spacing, and layout at the minimum window size.
5
columns at most, from the minimum window size up to 1792 px and wider.

The same page at the minimum window size and at five columns.
Tags that don't hide information
Tags help users filter visually: they scan the cards and spot the add-ons relevant to them. Hiding tags behind a "…" tag would have broken that. Localization made it harder: in German, every tag is much longer, so a solution that only works in English wasn't enough. I designed a scrollable tag bar instead. It works for any number and length of tags, and the arrows show users that more tags are available: on the right at the start, on both sides in the middle, on the left at the end. When the tags fit, there are no arrows.

The tag bar when the tags fit, and in its three scroll states.
A detail page for what the card can't hold
The card answers "what is this?". The detail page answers "should I download it?": full description, all tags, version, file size, compatibility, publish date, and links to the publisher's support page and website. I left out download counts and ratings. Developers couldn't have built a rating system in time, and it wasn't needed in this context: users install add-ons they're already entitled to, so I didn't want to copy an app store. We also dropped breadcrumbs, because a back arrow was enough for so few levels. Whether an add-on is available to you shows only on the detail page, not on every card.

The detail page: the full description, all tags, and the facts you need before downloading.
04 — the result
Ready for development, never shipped
I delivered the card family for grid and list views, the tag logic, the detail page, the resize logic for every window size and a specification for developers. The project was stopped when another team started building a very similar product in parallel, and neither version shipped, so there's no usage data to report.
05 — what I learned
Solve the problem users actually have
Check what users are asking for
We spent half the project on one-click install, a problem nobody had asked us to solve. A much simpler catalog covered exactly what users needed. What users already ask for should decide how big the solution is.
Build on what the product already has
Studying Project Hub first meant the store felt like part of Archicad, and I didn't re-solve problems that were already solved.
Design for the extremes
The longest name, the most tags and the narrowest window decide whether a card works. The average case almost always does.
Narrow directions earlier
Keeping many options open felt safe, but it made reviews harder. Now I name the decision each variant is meant to settle.