JST-2 generates a 3D scene you can enter and keep expanding, with characters and motion held in the same representation. It is for creators who already animate characters in Viggle, and for developers building games, interactive scenes, and simulation environments. The launch matters now because many world models still learn physical structure indirectly from pixels and video. Viggle spent four years on Joint Space-and-Time so position, appearance, and change over time are represented directly in 3D. The model treats a scene as persistent atoms. Each atom is a piece of matter or an object part that keeps its identity as it moves or deforms, with an explicit xyz position and learned vectors for attributes and dynamics, encoded as Physics-Native Tokens. A feed-forward pass produces the 3D representation directly, without per-scene optimization. Viggle reports about one second of model inference on an NVIDIA B200, excluding network latency, at a cost below $0.01. The decoder turns those tokens into images or 3D Gaussian splats for browser viewing. New viewpoints can be filled in as you move, with up to 64 input views held as spatial context, and a scene already generated can be explored again without another model run. A character can come from a single image, including a cropped or partly hidden one, with motion dynamics learned into the tokens so there is no separate rigging and skinning step. Text or video can drive full-body motion. In blind tests Viggle published, raters preferred these characters to Meshy 7.1 and Tripo H3.1, with the strongest results on facial detail, hand articulation, and how the body and clothing deform. This extends JST-1, the character model behind Viggle's meme app, which the company says has reached more than 50 million users. The Toronto team, led by co-founder and CEO Hang Chu, developed JST-2 from 2024 to 2026. It is available in the browser through PINOC and through the API, including from coding agents. The same model can also reconstruct a complete 3D environment from sparse real-world photos, which Viggle points toward simulation, mapping, and related spatial work.