Google has started rolling out Project Genie, an experimental AI-powered prototype that allows users to create, explore, and remix interactive virtual worlds using generative artificial intelligence. Developed under Google Labs and powered by the advanced Genie 3 world model, Project Genie represents a significant step forward in AI-driven immersive experiences and real-time world simulation.
At present, Project Genie is available to Google AI Ultra subscribers aged 18 and above in the United States, with access being introduced gradually. The launch follows limited early testing and aligns with Google’s broader push into generative AI, spatial computing, and interactive digital environments.
Project Genie is a web-based AI research prototype that gives users direct access to Google’s general-purpose world model. Unlike conventional 3D creation tools that generate static scenes, Project Genie produces dynamic environments that evolve continuously in response to user actions and movement. The platform is powered by a combination of Google’s Genie 3 world model, Gemini AI, and Nano Banana Pro, which provides fine-grained visual control and image refinement. Google first previewed Genie 3 in August 2025, highlighting its ability to generate fully interactive worlds rather than fixed 3D environments, and Project Genie now brings that research into a usable experimental experience.
Google explains that Project Genie is built around three core capabilities: world sketching, world exploration, and world remixing. Through world sketching, users can create AI-generated environments using text prompts, generated visuals, or uploaded images. They can design landscapes, cities, or fantasy settings, define characters, choose modes of movement such as walking, driving, or flying, and select first-person or third-person perspectives. Integration with Nano Banana Pro allows users to preview, refine, and adjust visuals before entering the generated world, offering greater creative precision.
World exploration enables users to navigate these environments as they are generated in real time. As users move through a world, the AI predicts and constructs the path ahead dynamically, with environments responding instantly to actions and camera angles adjustable during exploration. This real-time responsiveness makes Project Genie feel more like a living simulation than a pre-rendered digital space.
With world remixing, users can build on existing environments by modifying their underlying prompts. They can remix their own creations or curated worlds from a gallery, draw inspiration from a built-in randomiser, and combine elements from multiple worlds to create entirely new experiences. Completed worlds and explorations can be downloaded as videos for sharing or documentation.
Project Genie is currently accessible via the web through Google Labs for eligible Google AI Ultra subscribers in the US. While Google has not yet announced a timeline for a global rollout, broader availability is expected as the prototype continues to evolve.
Overall, Project Genie underscores Google’s increasing focus on generative AI for immersive experiences, AI-powered simulation and spatial intelligence, and creative tools for gaming, education, research, and virtual design. By enabling real-time, AI-generated interactive worlds, Project Genie has the potential to influence the future of game development, virtual training, digital storytelling, and next-generation metaverse-style applications.
