We build at the intersection of spatial computing, AI intelligence, and development automation — tools and experiences that didn't exist until now.
We design and build immersive spatial applications — from WebAR that runs in a browser to full VR environments. Gesture-driven, gesture-aware, and built for the hardware your users actually own.
Your features fall off, one by one. Catch each with a tap — or a blink — and watch your face come alive, scrambled and still recording every move. Runs in the browser, installs like an app.
Play now → Live · Gesture-controlledFly a hand-drawn character with your hands alone — swoop, dodge quadrant by quadrant, and chain gestures through four shifting dimensions. No controller, no install, just your camera.
Play now →Browser-native augmented reality using MediaPipe hand tracking and Three.js rendering. No app install required — your users just open a link.
WebXR · Three.js · MediaPipeOn-device ML gesture recognition with sub-frame latency. Designed for low-end hardware — if it runs on a Galaxy A03 Core, it runs everywhere.
MediaPipe · On-device MLProcedural dungeon generation, PBR materials, dynamic lighting, and particle systems — all running in WebGL at smooth framerates on mobile browsers.
Three.js · GLSL · WebGLSpatial experiences optimised for the devices most people carry. We target mid-range Android as a baseline — because democratised XR is better XR.
Android · iOS · ProgressiveWe automate the friction out of software delivery — from incident triage to deployment pipelines. Every tool we ship follows one rule: reduce manual steps to zero.
Alert fires → Jira ticket created → intelligence pre-filled → strike team assembled. No manual steps, no context switching.
Runs inside Atlassian's security perimeter. Your data never leaves. Deployed via Atlassian Marketplace with one click.
P1 created → tracking issue auto-opened in your repo with AI analysis in the body. First on the Atlassian Marketplace to do this.
Register an agent endpoint and let AI resolve issues autonomously — with human approval gates and full audit trail.
One click from incident to war room. Multi-domain experts invited automatically. AI analysis posted as the opening message.
Sarah Okonkwo
4 similar resolutions · Platform EngA quick translation, because this one lives a few layers down. Most of what we make is software you can open in a browser. Stage Zero was the opposite — a hardware research track. It looked at the physical foundation that high-end chips are built on: not the processor you'd recognise on a spec sheet, but the groundwork underneath it that quietly decides how fast the next generation is even allowed to go.
We weren't designing a chip. We were trying to call a direction — the way the most powerful processors would have to be packed tighter and wired differently to keep getting faster. The kind of shift you only see clearly once a flagship graphics card or AI accelerator ships it, and the rest of the industry spends two years calling it obvious.
By the time our read was clean, that future had already turned up in shipping silicon. The top-tier parts were moving in exactly the direction we'd sketched — which, for a research bet like this, is the best outcome there is. You don't keep proving a thesis the market has already cast in hardware. So we shelved Stage Zero, kept what it taught us, and started building one layer up.
Questions about Ticket Intelligence, a custom AR/VR project, or something entirely new? We read every message.