Overview of Project Genie
Project Genie is an experimental research prototype that enables users to build and explore interactive worlds using natural language and optional images.
Eligibility and Access
The service is currently limited to Google AI Ultra Subscribers located in the United States who are over 18 years old, with plans to broaden availability in the future.
Crafting Environment Prompts
Effective prompts start by stating the desired setting-such as a forest, city, or lunar surface-and then layer specific details like weather, objects, and visual style. Use concise, declarative sentences to describe attributes like lush or wintry for a forest, or stormy for a city.
Designing Navigable Characters
Choose a character that serves as the user's avatar within the world. The character can be any entity-a giraffe, robot, or pixelated doll-and its movement mode (fly, roll, run backward) should be explicitly described. Adding a visual effect, such as a puff of smoke when the character hops, enriches the experience.
Incorporating Images
Users may upload an image to define the environment instead of starting with text alone. The image should feature the character centered in the frame with sufficient background to convey the surrounding setting. Supplement the upload with textual details to refine lighting, weather, or material qualities.
Choosing Perspectives and Interaction
After the world and character are created, select a navigation mode: first‑person view to see through the characters eyes, or third‑person view to watch the character from above. The perspective button toggles between these options, allowing real‑time exploration and iterative prompt adjustments.