Initial problem
Most augmented-reality products hardcode one experience directly into one app. Social posts, games, navigation, and brand activations each become separate rendering and backend systems, while external creators either receive excessive device access or cannot participate at all. A reusable platform needed one spatial model, a controlled capability boundary, and server-authoritative state without reducing the native AR experience to embedded web content.
Implementation
Beyond is a native AR runtime that loads activatable Reality Layers around the viewer. Every visible element is a Spatial Object owned by a Layer, whether it is a social Drop, quest, place, navigation marker, or partner experience. The iOS runtime uses ARKit and RealityKit; a Go control plane resolves scenes, permissions, assets, moderation, and server acknowledgements. First-party social content uses the same manifest and Layer API intended for partners, while a capability broker provides minimized camera, location, and identity access instead of exposing device frameworks directly.
Outcome
My contribution
Impact
What mattered
One Layer API for first-party social content and future external partners
Generic Spatial Object model for Drops, quests, places, navigation, and brand moments
Native iOS runtime using ARKit and RealityKit instead of an embedded web scene
Capability broker that minimizes access to camera, location, identity, and interactions
Server-authoritative interaction acknowledgements, moderation states, and kill switches
Deployed core loop from sign-in and scene query through object creation and realtime updates
Insights
The Berlin field preview visualizes active layers, anchors, spatial objects, and runtime status
Server acknowledgements, the capability broker, and layer kill switches form the platform safety boundary
Technical frame
The technical frame sits at the end on purpose: it shows the engineering substance behind the solution.
Next project
For the moments that stay between us.
View project