Runway Labs Previews Project Continuum, an AI-Generated Video Operating System

Runway Labs, Runway’s research division, used a demo video and social posts on October 1, 2026 to introduce Project Continuum, which it calls “a new operating system research application built around real-time video interfaces.” Rather than drawing windows, panels, and menus with traditional code, the system generates them frame by frame using a video model, continuing the ideas behind Runway’s earlier Solaris research.

Details

• What it is: a research prototype, internally labeled “PROJECT CONTINUUM® CN-01 | REAL-TIME VIDEO OPERATING SYSTEM | v0.0.1-ALPHA,” that generates an entire OS-style interface live rather than rendering it from HTML, CSS, or a DOM
• Portals: generated video panels spawned from on-screen selections that persist and regenerate when resized, instead of recalculating a layout the way a browser would
• Visual Thinking: AI agents render their own reasoning as real-time 3D visualizations that users can iterate on directly
• Responsive Video Interfaces: the system generates interface suggestions on the fly based on whatever content is active, rather than relying on fixed, pre-built layouts
• Interactive Worlds: full, streamed 3D environments — Runway’s demo includes a Mars outpost — that users can explore and interact with in real time
• Availability: Runway was explicit that none of this is a product yet — there is no sign-up, waitlist, request form, pricing, or API, and the company has not announced a release date
• The reveal followed Runway’s first AI Summit in San Francisco, which closed on September 30, 2026, where the company had separately introduced its Praxis-1 robotics model the day before

What happened next

Runway framed the reveal as deliberately incomplete, closing its announcement with “that’s just an early look” and a promise to share more in the future. Project Continuum signals that Runway’s ambitions now extend well beyond video clip generation toward reimagining how entire software interfaces could be produced by generative video models, a research direction that, if it ever ships, would put Runway in more direct competition with how operating systems and app interfaces are built today rather than just how video and images are created.