RealityKIT
AI vision for business — inside a headset. Mixed reality for industrial work: overlays, instructions, and perception for Quest 3 and HoloLens 2.
RealityKIT is a mixed-reality runtime built for operators — technicians, pickers, inspectors, surgeons — who need contextual data pinned to the physical world. It targets Meta Quest 3 and 3S passthrough and Microsoft HoloLens 2, and its job is to close the loop between what a camera sees and what a person does next. Where VisionKIT is AI eyes on the server, RealityKIT is AI eyes on the head.
A Unity core targets OpenXR with the Meta XR SDK for Quest passthrough and MRTK3 for HoloLens 2. Hand and eye tracking drive interaction; spatial anchors and scene understanding keep overlays attached to real objects across sessions. Perception — detection, OCR, classification — is shared with VisionKIT and served either from an edge box over WebRTC or low-latency gRPC, or compiled on-device with ONNX for air-gapped sites where latency rules.
On day 1 you get a passthrough scene on Quest 3 or HoloLens 2 with live detection and OCR overlays pinned to real objects, plus a remote-expert channel.
By week 3 the perception models are fine-tuned on your equipment and workflows, the overlays reflect your SOPs and manuals, and builds distribute to your fleet via Quest for Business or HoloLens device management.
Build it from scratch, or start a week ahead
- Six to nine months building the orchestration, memory and recovery before a first real result.
- An evaluation harness you write from scratch — then argue about.
- A team learning your edge cases live, in production.
- A black-box vendor you can't inspect, tune or move off.
- A working RealityKIT on your data in week one — the hard parts already solved.
- Evals seeded on day one and graded against your real workflows.
- Senior owners with full traces and dashboards from the first deploy.
- You own the prompts, the weights, the traces and the outcomes.
Ships with the hard parts solved
Passthrough MR runtime
Unity + OpenXR targeting Quest 3 / 3S and HoloLens 2, with hand and eye tracking.
Perception overlays
Detection, OCR, and classification from VisionKIT rendered in-world on real objects.
Spatial anchors & scene
Persistent anchors and scene understanding so overlays survive between sessions.
Remote expert & capture
Live video + annotation channel so an off-site expert can mark what the operator sees.
Where teams deploy it
Fine-tuned on your data and shaped to the workflow it lands in — these are the deployments we see most.
Kick-off to production in three weeks
A fixed scope and a visible finish line — you see it working before it's load-bearing.
Week 1 — Integration
Device targets, passthrough scene, and VisionKIT inference link wired in.
Week 2 — Fine-tune
Perception models and overlays tuned to your equipment and SOPs.
Week 3 — Ship
Distributed via Quest for Business or HoloLens MDM to pilot operators.
Most MR pilots stall because the interesting part — the AI that makes the overlay useful — was an afterthought. RealityKIT starts from perception and treats the headset as a display for it, so the pilot actually saves minutes on a task instead of being a novelty demo.
“They understand our needs quickly and are a delight to work with.”
RealityKIT, live on your stack in a week
One call to scope it, a senior team on it from day one — and a fixed scope agreed before we start.