Runway Unveils Project Continuum, an OS Research Project Built Around Real-Time Video Interfaces
On October 1, Runway Labs released the research project Project Continuum, officially framed as an early exploration of an operating system built around real-time video interfaces. It proposes a new interaction model of "meeting you where you are" and demonstrates four new ways of interacting with computers. What makes the project notable is that it pushes real-time video models beyond mere content generation into an interface- and OS-level interaction paradigm.
Confirmed
- Project Continuum is released by Runway Labs, built around real-time video interfaces, and is early-stage OS-level research
- Portals: Generated content appears with soft, dreamlike edges that clearly mark the boundaries of AI-generated areas; Portal windows can be dragged onto the page and coexist side by side with the user's existing content
- Visual Thinking: A new harness for coding agents that uses a real-time video model to dynamically visualize every step and tool call an agent makes, keeping developers "in the loop" while complex autonomous workflows run
- Expanding any media into a real-time world: Images, videos, and other media can be expanded into interactive real-time world simulations, turning fixed experiences into immersive multimodal simulated environments for humans and agents; the demo showed spatially aware reconstruction and real-time click-based interaction
- Scalable interface: The video interface can be scaled freely, with the model recomposing the scene in real time with spatial awareness, and users can interact directly by clicking inside the frame
Why it matters
- The project combines real-time video models with the OS interaction layer, exploring new ways to fuse generated content with the user's workspace rather than existing merely as a generation tool
- Visual Thinking targets observability and human-in-the-loop concerns for AI agents, addressing real needs around debugging and supervising autonomous workflows
- Portals' "dreamlike edges" design shows the team deliberately distinguishing AI-generated content from user content at the interface level, touching on issues of trust and identifiability
2026-10-01 ~ 2026-10-02 · 6 related posts
Primary sources
- Runway Unveils Project Continuum, an OS Research App Built on Real-Time Video Interfaces — runwayml ·
- Runway launches Visual Thinking: real-time video model that visualizes coding agent workflows — runwayml ·
- Runway Labs unveils Project Continuum, an OS research app built on real-time video interfaces — runwayml ·
- [source] Runway Unveils Project Continuum, an OS Research App Built on Real-Time Video Interfaces — runwayml · 2026-10-01
- Runway Details Portals: Draggable Windows of Generated Content With Dream-Like Edges — runwayml · 2026-10-01
- Runway's Video Interfaces Resize Freely as the Model Recomposes Scenes in Real Time — runwayml · 2026-10-01
- [source] Runway launches Visual Thinking: real-time video model that visualizes coding agent workflows — runwayml · 2026-10-01
- Runway demo: expand any media into an interactive real-time world simulation — runwayml · 2026-10-01
1 near-duplicate retellings: runwayml