Skip to main content
Demo · In the standard

Telepresence.
A call is a live world of one avatar.

This page is a live skeletal world. Your tab is a signing actor: it creates its own three-bone arm rig and animates it with signed SET_ROTATION ops, fanned out over the same relay the multiplayer worlds demo uses (or a BroadcastChannel between tabs as a fallback). Every sink replays the same signed op set to the same pose hash, so the tabs converge — poses agree @ tick N. Open this URL in a second tab, or on another device, and a second avatar joins the room and waves back. The motion is signed and auditable; the appearance — a skinned mesh — is each sink's to render.

● connecting…
avatars — convergence — tick — pose hash —

Each tab is an actor

A 32-byte key is minted in the tab and never leaves it — only signed ops do. Each op is a CausalOp a native sink would ingest, identical to the worlds path; the relay verifies every signature at the door and fans it out, adjudicating nothing.

Pose-equivalence converges

There is no central truth — each sink replays the combined op log and computes the same BLAKE3 pose hash at every boundary. When two tabs agree, the body's motion is cryptographically identical on both, with no server holding the canonical frame.

The wire is the skeleton

What travels is signed joint rotations, not pixels. The sink is a computer: it holds the rig and renders the body. The same machinery that runs a shared physics world runs a shared body; telepresence is a live world whose nodes are bones.

The avatars and interactive-worlds extensions, the reference engine and the relay this page uses are all in the open-standards repo under Apache-2.0. Transport (relay or BroadcastChannel) is outside the extension's scope; what travels is exactly the signed op a native sink ingests. The skeleton drawn here is the verified pose; a sink with a rigged mesh skins it. Energy: not metered in this demo.