A printed card as the starting point for an interactive 3D experience — a one-week exploration of WebAR, interaction, and what happens when physical objects become part of the interface.
The brief was open-ended: build something creative around immersive design. We wanted to explore interaction beyond the screen, starting from a simple question: What if a printed card could become the gateway to a 3D experience that can't physically fit inside it?
Three inventors. Three inventions. One physical-to-digital interaction. Scan an inventor, find their invention, and bring it to life in AR.
Printed inventor and invention cards become the entry point into the experience.
MindAR recognises the cards through the device camera, with no app installation required.
Three.js renders and animates each invention as it appears in the user's space.
Audio accompanies each reveal, while a minimizable information card keeps context available without competing with the AR object.
Four moments shaped Invento — from a simple scan-and-reveal prototype to a tested and more deliberate AR interaction.
With one week to go from idea to a working WebAR experience, the first prototype stayed deliberately narrow: scan an inventor card, and the invention appears in AR with information and audio. The challenge wasn't getting a 3D model on screen — it was making the interaction understandable, responsive, and enjoyable.
The first version worked, but the reveal felt too passive. We added a matching layer: the user identifies the correct invention before seeing it come to life, and a wrong match gives a hint instead of a dead end.
Once the prototype worked, we put it in people's hands and watched what they actually did. Five problems surfaced:
The info card and audio could appear before the 3D model had finished loading, making the reveal feel broken.
Rather than trying to speed up the model itself, we preloaded all three invention models when the experience starts. By the time a match is found, the reveal is ready.
Each observation became a specific interface change — shown here against the screens it came from.
Scanning order wasn't obvious and people tried both cards at once.
One instruction at a time: "Scan Inventor," then "Scan Invention."
Instructions, progress and AR all competed for attention, making it cluttered.
Simplified interface — only what's needed for the current step.
Audio and information appeared before the 3D model was ready.
Models are preloaded and ready when the match is confirmed.
We compared the early and refined versions across three measures to see what actually changed.
From a successful match to a visible 3D model.
Reached the AR reveal without assistance.
Inventor → invention, without additional explanation.
The camera can recognise the card, the model can load correctly, the AR can fire exactly as intended — and a user can still wonder what to do next. Testing shifted our focus from making the technology work to making the experience make sense.
Moving a camera, holding a card steady, searching through a pile — in an immersive interface, these count as interaction too, not just a prelude to the "real" screen.