Spatial construction, scale, materials, optimization and preparation before the asset enters the web pipeline.
A browser-based spatial system for turning rooms, properties and designed environments into portable AR experiences. I developed the concept from architectural-scale experiments into a repeatable web-to-mobile pipeline that lets someone preview a space, launch it from a link or QR code, and step into it from their own device.
ENTER ORB WALK-IN
Orb Walk-In grew out of experiments I was already doing around launching architecture into augmented reality. I was interested in a simple problem: photos and video can show a room, but they still keep the viewer outside of it. I wanted the environment itself to become portable, something a person could call up on a phone and experience spatially without installing a dedicated application.
The orb became the interface between those two states. From the outside it behaves like a contained preview of another place. Once launched in AR, the scale relationship changes and the viewer can physically approach, inspect and step into the environment. That made the idea useful beyond a single demo: real estate, interiors, hospitality, exhibitions and other spatial experiences could all use the same underlying system while changing the world inside it.
Orb Walk-In had to work even when the link was opened inside platforms that interfere with a normal Model Viewer or AR launch. I built the entry flow so the experience can recognize that friction and reroute the visitor to a clean external landing page instead of allowing the interaction to fail.
From there the experience can hand off cleanly between devices. Desktop users receive a QR bridge to mobile, while phone users can inspect and scale the environment before launching it into AR. The result is a distribution layer designed around the reality that the first click may happen almost anywhere.
A normal web link is the entry point. Orb Walk-In can be shared through a property page, campaign, social platform or direct message without asking the visitor to understand the delivery system behind it.
The core interaction is a change of scale. The orb can first be understood as an object, but the experience becomes meaningful when the environment is placed at a scale that relates to the viewer's body. A room stops being an image of a room and becomes a spatial reference the person can move toward and around.
That transition is why I kept the launch process lightweight. The technical work sits behind an interaction that should feel almost ordinary: open, place, walk. The demonstration shows that shift directly, from handling the spatial object on a phone to stepping into the environment at human scale.
Spatial work becomes a performance problem very quickly. A model can look correct in Blender and still be the wrong asset for a mobile browser. I had to think about geometry, textures, file size, materials, scale and how the same environment would behave when it moved from a 3D workspace into web delivery and then into AR.
I built the workflow around reusable stages rather than treating every room as a new technical experiment. Source material is prepared in Blender, optimized for runtime, exported into the formats required by the target devices, tested in the browser, and then iterated against real mobile behavior. That pipeline became part of the product itself.
The spatial environment is assembled and prepared as a real 3D asset, with scale, materials and geometry considered for the runtime it will eventually enter.
Orb Walk-In does not need a large application shell to make the spatial idea work. The technical system is deliberately lightweight: optimized 3D assets, browser presentation, device-aware AR launch paths and static hosting are combined into a delivery system that can be linked from an ordinary website.
This project is where the granular implementation details matter. Unlike Emory, where the strongest story is consultation and cross-platform client delivery, Orb Walk-In is my own system. The case study can expose the actual assembly: how the assets are built, converted, hosted, launched and tested, and how each layer contributes to making the final interaction feel simple.
Spatial construction, scale, materials, optimization and preparation before the asset enters the web pipeline.
A lightweight browser layer for presenting the 3D environment and bridging compatible devices into AR.
Device-appropriate spatial formats supporting browser preview and mobile AR delivery across Android and iPhone.
The interaction and presentation layer around the spatial asset without requiring a dedicated native application.
Fast static deployment for testing, iteration and public access to the browser-based experience.
Asset delivery and hosting infrastructure that keeps larger spatial and media files separate from the page shell.
The earliest builds were full walkthrough models aimed at maximum realism. They proved the spatial premise, but they also revealed the tradeoff between visual detail and the portability required for fast browser and mobile AR.
The interesting part of Orb Walk-In is not only that a room appears in augmented reality. It is the set of decisions required to make that room portable: how I model and optimize it, which formats are necessary, how the browser decides what to launch, where the assets live, and how the same experience is tested across devices.
The build moved repeatedly between authoring, conversion, browser tests and real-device tests. GLB assets had to remain lightweight for the web, while iPhone delivery required USDZ versions. Depending on the asset, I converted those files with online tools or used Reality Composer when animation needed to carry across into the Apple AR version.