Google has started rolling out Project Genie, an experimental research prototype that allows users to create, explore, and remix interactive virtual worlds. Developed by Google DeepMind and hosted on Google Labs, Project Genie is an early-stage tool with several limitations, including a 60-second world generation cap. Here’s everything you need to know about Google’s ambitious new world-generation project.
Project Genie is a general-purpose world model developed by Google DeepMind and built on Genie 3. The tool is designed to generate and simulate interactive environments in real time, allowing users to build virtual worlds that evolve based on user actions.
Hosted on Google Labs, Project Genie enables real-time world creation and exploration. A world model explains how an environment reacts and changes over time and predicts how user actions affect outcomes.
Earlier AI systems from Google DeepMind were trained on stable environments such as chess and Go. However, building more general AI systems requires models that can operate under diverse and changing conditions. Genie 3 addresses this requirement by dynamically generating environments based on continuous user interaction.
Project Genie is available starting today in the United States for Google AI Ultra subscribers aged 18 and above. Google has confirmed that the tool will roll out to additional regions over time, with continuous updates and improvements planned.
Since Project Genie is still an early-stage research prototype, it comes with several limitations:
World generation is currently limited to 60 seconds
Generated environments may not always accurately reflect real-world physics
The tool may not always fully follow user prompts
Google states that these limitations will be addressed gradually through future updates.
Project Genie is a web-based prototype powered by Genie 3, Nano Banana Pro, and Gemini, and it focuses on three core functions:
Users can generate worlds using text prompts, along with generated or uploaded images. The tool allows users to create:
Characters
Environments
Movements and actions such as walking, riding, flying, and driving
This feature is integrated with Nano Banana Pro, enabling more flexible and creative world-building.
Once a world is created, users can fully navigate the environment. As users move forward in the space, Project Genie dynamically generates new paths and surroundings in real time based on user actions. Camera positions can also be adjusted during exploration.
Users can remix existing worlds by adding new elements or modifying current prompts. The prototype includes:
Curated worlds for exploration
A randomizer gallery for experimentation
Users can also download videos capturing their worlds and exploration sessions.
Genie 3 brings physics-aware interactions, object behavior, and environmental stability, making Project Genie useful across multiple fields, including:
Robotics research
Animation and visual storytelling
Fiction and creative world-building
3D modeling
Historical simulations and research
The project is based on insights gathered from large testers across multiple industries and sectors.