Reality Layer Runtime / Berlin Field Preview

Beyond

The world has layers.

Beyond loads activatable Reality Layers exactly where they exist: Drops, Quests, Places, and partner objects rendered as Spatial Objects in one server-authoritative runtime.

Join the Field Test Developer Access

01

Layer API

Social is only the first Layer. First-party Drops use the same public API external partners get: manifest in, Spatial Objects out.

{
  "id": "com.beyond.social",
  "schema": "beyond.layer.manifest.v1",
  "objects": "spatial_objects",
  "capabilities": ["location:coarse"],
  "interaction": "server_ack"
}

com.beyond.social drop.created
com.partner.venue scene.enter
com.partner.creaturehunt interaction.ack
com.beyond.nearby object.updated
com.beyond.debug layer.published

02

Spatial Objects

Drops, Quests, brand moments, and navigation arrows all live in one generic spatial model. No per-use-case tables, ever.

anchor.geo
anchor.plane
asset.glb
component.tap
status.approved

03

Scene Query

The runtime asks for a Scene around the viewer. The backend filters by status, distance, capabilities, and moderation before anything renders.

layer.status
object.status
distance.h3
capability
moderation
visibility

scene.fetch 24ms
filter.status ok
compose 12 obj
render 60fps

04

Server Ack

Clients report interactions; the backend or Layer server decides state. captured = true never originates on the client.

05

Capability Broker

Layers never touch camera, ARKit, or location directly. Declared, approved grants hand out minimized data — nothing more.

06

Kill Switch

Every Layer and object carries a status. Killed Layers vanish from the next Scene fetch and are pushed out over realtime.