Rockland Technologies
About Us
We are a small team in the Bay Area. We build tools for making, sharing, and viewing 3D digital assets. This is how we got here.
A switch on a wall
We started in 2024 doing lighting controls work in New York City. Commissioning buildings, programming systems, designing and making user interfaces with architects and designers. We learned a lot about how people control the space around them, and we kept hitting the same wall. Every good idea turned into another panel on a wall or another app on a phone. You still had to walk somewhere, or find the right screen, before anything happened. Cold Brew Lighting came out of that work.
Something clicked
When Apple Vision Pro was released, something clicked. If the headset already knows what you are looking at, then looking at a light should be enough to turn it on. No panel. No menu. No hunting. So we built Air Orbe, an app that lets you control smart devices just by looking at them.
The bulb went on
Air Orbe worked. What surprised us was everything around it. Spending real time in this medium made it obvious how much was still missing and how much was possible, and we did not want to do it on the side. We packed up, drove across the country, and moved to Silicon Valley. An apartment and an office. We kept going on Air Orbe for a while, then admitted something to ourselves: controlling lights was not the problem worth the next few years of our lives. The bigger one was sitting right in front of us. There was still no good way to create, share, and view 3D things across the devices people already own. So we started AirVis. The first version was a place for digital objects and animations, mostly USDZ files, on Mac, iPhone, iPad, and Vision Pro, so something you made on one device opened cleanly on the others.
Chasing better light
Around that time NeRFs and Gaussian Splats were picking up. These captures feel like the place you were standing in, not like a model of it, and we wanted AirVis to handle them. That is where the problem got much bigger than anything we had solved before. Nothing we could pull off the shelf held the quality we wanted on the devices we cared about. So we wrote our own rendering engine from scratch, in Metal, by referring to open source projects like MetalSplatter.
Handing over the switch
Viewing was only half of it. If you cannot make a capture yourself, none of it matters. So we built a trainer by referring to open source projects like gsplat and got it working end to end. Scan with your phone, then walk through the result in VR. Then we brought it everywhere else we could: Android, Meta, Windows.
Turning it on for everyone
One thing still bothered us. Training costs money, and there is a real learning curve to making a good capture. Anyone curious had to pay before they knew whether they even liked it. We wanted the cheapest possible way in, we provided free captures but with daily limits and capture time limits, but that was not enough. So we built a local solution that runs on the machine you already have, not just an expensive one. That became AirVis Studio, and that is where we are today.
Many small little lights along the way
We focused so much on rendering that the engine can now handle hundreds of millions of splats using level of detail, and we built a foveated streaming solution on CloudXR so you can view a capture at full quality in VR with an NVIDIA GPU carrying the load. We worked with World Labs on quality checks for Spark. And a fair amount more, just lazy to write all that down.