Editorial illustration for Runway's Solaris Previews No-Code Internet With OpenClaw 2.0
Runway's Solaris: No-Code Web Apps Built With AI Video
Runway's Solaris Previews No-Code Internet With OpenClaw 2.0
Runway put out a preview this week of something it calls Solaris, an "Interface World Model" that skips code entirely and renders websites and apps as live video, frame by frame, as someone clicks and drags. There's no HTML underneath, no backend logic running. Instead a video model generates each screen on the fly, guessing at what should appear next based on the action just taken.
The company built Solaris by pairing its Gen-4.5 video model with an LLM that interprets user input, a click, a drag, a scroll, and turns that into a prompt for the next frame. Runway showed off a shirt getting dragged from a rack onto a photo, a salad assembled ingredient by ingredient, and a small combustion simulation, all generated in real time with nothing resembling traditional code behind them.
Runway also ran its own comparison, pitting Solaris against interfaces built by Claude Opus 5. The results, and the rough edges the company copped to, point to how far this kind of interface generation still has to go before early access users get their hands on it.
In a Runway study, testers picked Solaris over pages coded by Claude Opus 5 in 71% of matchups judging in-scene behavior and 61% on instruction following.
Why this matters
Solaris is Runway betting that the browser as we know it becomes optional. If an interface can be generated frame by frame from a prompt, the whole idea of "building a website" or "shipping an app" starts to look quaint. That's a big claim, and we'd want to see load times, consistency across sessions, and actual production use before calling it real. Rendering the internet as continuous video is a neat trick; running a business on it is another matter.
The same week's other releases point the same direction. OpenClaw 2.0 going multiplayer means agents aren't just assistants anymore, they're coordinating with each other. Fireworks letting teams train their own open models on proprietary work signals the same shift: less renting intelligence, more owning it.
For developers and founders, the practical move right now isn't chasing the flashiest demo. It's figuring out which of these tools you can actually build on this quarter, and which ones are still previews dressed up as products. Watch what ships, not what streams.
Related Reading
LLMs & Generative AI ChatGPT's 'Nerdy' tweak rewards goblin metaphors in answers, study finds
LLMs & Generative AI Google tests visual 'magazine-style' UI for Gemini 3 Pro users
LLMs & Generative AI AI Engineers Face Rising Costs, Need New Strategies for Efficiency
From Archives Nine LLMs Tested on 2,005 Oncology Decisions in New Benchmark
From Archives Google's TimesFM-3 Tops Time Series Benchmarks
Common Questions Answered
How does Runway's Solaris generate websites and apps without traditional code?
Solaris uses an Interface World Model that pairs Runway's Gen-4.5 video model with a language model to interpret user input like clicks and drags. Instead of rendering HTML with backend logic, the video model generates each screen frame by frame on the fly, predicting what should appear next based on the user's actions.
What were the results of Runway's comparison study between Solaris and Claude Opus 5?
In Runway's testing, Solaris was preferred over pages coded by Claude Opus 5 in 71% of matchups when judging in-scene behavior and 61% of matchups on instruction following. These results suggest that the AI-generated video interface performs competitively against traditionally coded solutions.
What does Runway believe Solaris means for the future of web browsers and app development?
Runway is betting that the traditional browser becomes optional as interfaces can be generated frame by frame from prompts, potentially making the concepts of 'building a website' or 'shipping an app' obsolete. However, the company acknowledges that production viability depends on factors like load times, consistency across sessions, and real-world business use cases.
What is an Interface World Model and how does it work in Solaris?
An Interface World Model is Solaris's core technology that renders websites and apps as live video rather than traditional code. It combines a video generation model with language understanding to interpret user interactions and continuously generate the next visual frame, eliminating the need for HTML and backend logic.
Further Reading
- Runway News | Introducing Solaris - Runway
- Runway's Solaris Renders Interactive Apps as Live Video, No Code Needed - AlphaSignal
- Runway's Solaris Turns Any Prompt Into a Functioning Interface, No Code Required - Glitchwire
- Everything That Happened in AI Today (Mon, August 31, 2026) - The Neuron
- All The Latest AI News In One Place - Future Tools from Matt Wolfe - Future Tools
