OpenArt has launched Worlds, a tool that lets you generate a persistent, navigable 3D environment from a single text prompt or image, walk through it using game-style controls, and capture production-ready shots from any angle you choose.
The feature is powered by World Labs, the spatial AI company co-founded by Stanford professor Fei-Fei Li, which raised $1 billion in February from backers including Autodesk ($200 million), NVIDIA, AMD, and Andreessen Horowitz. OpenArt is the first consumer-facing platform to integrate World Labs’ spatial AI into a full creative suite.
The practical significance is straightforward. Until now, AI video creators working on anything that required a consistent location across multiple shots – a recurring set, a branded environment, a character’s apartment – had to rely on workarounds. You’d regenerate backgrounds and hope for consistency, or spend hours in post-production aligning frames that the AI had decided should look slightly different each time. OpenArt Worlds solves this by building the environment once as a navigable 3D space. The world persists in your library. You return to it whenever you need another shot, and the lighting, geometry, and spatial relationships stay the same.
Once inside a world, creators can adjust focal length (from wide 23mm to telephoto 300mm), set aspect ratios, and position characters or objects into the scene using prompts or saved character sheets. The captured image then serves as a keyframe for video generation through integrated tools including Kling and Veo3, or through the platform’s own motion sync and lip sync features.
Future versions will add 3D export in formats including Gaussian splats (a format that represents 3D scenes as collections of overlapping soft blobs rather than traditional solid surfaces) and 3D mesh, opening a route into professional tools like Blender, Unreal Engine, and Cinema 4D. That capability isn’t live yet, but it’s on the roadmap.
Pricing starts at $14 per month on the Essential plan, rising to $29 (Advanced) and $56 (Infinite), all using a credit-based system. Annual billing brings the entry point down to roughly $7 per month.
World Labs also operates its own tool, Marble, which generates navigable worlds from single images. But OpenArt’s advantage is integration: Worlds sits inside a broader creative suite that includes image generation, video generation, character training, and audio synchronisation. For creators who want to build a world and then produce content from within it without switching platforms, that’s a meaningful distinction.
The quality ceiling is real. Detail degrades at distance from the central generation point, proportional errors crop up (doors that are too narrow, objects that don’t quite fit the space), and the environments aren’t yet photorealistic. Creators working at the top end of production quality will still need post-production refinement. But for the vast majority of AI video workflows – social content, branded series, commercial previs, and serialised storytelling – the tool is usable now.
OpenArt Worlds is available at openart.ai/suite/world.
