Google's Project Genie Opens Its Gates, But at a Premium Price

A year after its initial unveiling, Google’s ambitious AI world-building tool is now available to a broader audience. Project Genie, the polished public version of what was internally called Genie 3, is live—but with a significant catch. Access is restricted to subscribers of Google’s top-tier, most expensive AI service plan.

The technology allows users to generate interactive, explorable environments from nothing more than a photograph or a written description. While not a true 3D construct, Genie creates a responsive video simulation. A user can input commands to guide a character through the generated world, making it feel like a living, navigable space.

Genie 3’s key technical advance was a form of long-term memory, enabling the AI to maintain consistency within its creation for several minutes—a notable leap for the field. The current Project Genie builds on that foundation, integrating with Google’s newer AI models, including Nano Banana Pro and Gemini 3.

Although Google offers a library of pre-made worlds, the core appeal lies in spontaneous creation. A user can upload a picture of a fantasy castle or simply describe a neon-drenched cyberpunk alley, and Genie will attempt to build a playable scene around it. The release marks a strategic move by Google to monetize its most advanced generative AI research, placing a cutting-edge digital toy squarely in the hands of its highest-paying customers.