Runway Unveils Solaris. It Builds Clickable Worlds. The Interface Is The Invention, Not The Output.

Runway has introduced Solaris, which they describe as an interface world model. The demonstrated use case involves uploading a photo of yourself, then clicking and dragging a shirt from a virtual clothing rack onto your image to preview how it looks. No code is required to construct these interactive environments.

This demonstrates the principle of generative interfaces. The mechanism at work is direct manipulation of AI-produced content, where the model does not merely output a static image but produces an interactive space responsive to user actions. The mental model to absorb is that AI is shifting from producing artifacts to producing environments. That distinction is rather important, even if it sounds subtle to the uninitiated.

Runway, a company already known for AI video generation, is building Solaris. The clothing store example is their own demonstration of the technology.