AR and e-commerce, before it was normal
We thought there was a missing step between seeing a product in your room and actually buying it.
ARhove
Augmented reality for e-commerce — put a product in your room at its real size before you buy it
ARhove was an attempt to answer a simple e-commerce question: what will this thing actually look like in my space?
Instead of judging a product from photos and dimensions, shoppers could place it in their own room at its real size and see how it fits before ordering
the question photographs cannot answer
Online, a product is usually a set of photographs and a dimensions table. That tells you what the product looks like and how big it is, but not what either of those things means in your room.
ARhove put the product into the room instead. Point the phone at the floor or wall, place the product, and see it at its actual size next to the things that are already there.
That makes choices like size, colour and configuration much easier to judge.
what we built
We built the platform and a mobile MVP, then tested it through TestFlight.
The interesting part was making the same experience work across very different phones.
On devices with LiDAR, the depth sensor could reconstruct the room and give us much more information about the space. On devices without it, we took another route: the shopper photographed the product, the system generated a 3D model from those images, and that model could then be used in the same AR experience.
One path used hardware that only some phones had. The other was designed to make the experience available to everyone else.
Alongside the mobile app, we built a desktop 3D scene for loading and configuring products. Changes such as colour could be made directly in the scene without rebuilding the model, which made working with product variants much more practical.
let config = ARWorldTrackingConfiguration()
config.planeDetection = [.horizontal, .vertical]
config.environmentTexturing = .automatic
if ARWorldTrackingConfiguration.supportsSceneReconstruction(.mesh) {
config.sceneReconstruction = .mesh
}
arView.session.run(config, options: [])supportsSceneReconstruction(.mesh) is the LiDAR check. With it, the app could use scene reconstruction; without it, it fell back to plane detection.
how it's built
- Fittingthe AR scene: placed entities, gestures, coaching overlay
- CaptureUSDZ model + thumbnail + title
- Product · VariantREADME marks these in progress and TODO
where it stands
The MVP worked. We demonstrated it to potential customers and tested the idea, but nobody bought it.
Some of the stores we talked to already had AR solutions of their own, so there was no commercial deployment for ARhove.
That’s where it ended: a real product hypothesis, a working MVP, and a useful answer to the question of whether anyone actually wanted it.
ARhove is not an active product today.
From the journal