Project Genie Guide: How to Create and Explore AI Worlds
Project Genie gives eligible users an interface to sketch, explore and remix generated worlds. Here's how to write better prompts without confusing a research prototype with a full game editor.
Google launched Project Genie to U.S. AI Ultra users in January 2026, then announced broader global access for eligible $200 AI Ultra subscribers in May. It supports world sketching and exploration; the January prototype documented 60-second generation sessions, with features subject to updates.
01 / CONTEXT
What matters and why
Genie 3 is the research world model; Project Genie is Google's experimental product for people to experience it. These are related but not identical. In its initial release, the product combined Genie 3 with image generation and Gemini to help a user create a starting world, select a character and explore from a chosen perspective.
Availability has evolved during 2026. Google's May announcement expanded the eligible audience globally and introduced Street View-based geographic grounding for supported U.S. locations. Readers should verify current subscription, age and regional terms instead of following older U.S.-only instructions.
Start with a compact world premise, a distinctive landmark and a clear viewpoint. Treat sessions as exploration experiments, not production game levels.
02 / DEEP DIVE
Check current access before planning
Google's January post limited Project Genie to U.S. adults on a qualifying AI Ultra subscription. A later announcement expanded eligibility to $200 AI Ultra subscribers globally and added supported Street View features. Those updates demonstrate why access guides must carry an explicit review date.
Availability can depend on rollout stage, account location, age and subscription. The public prototype is a Google Labs experiment, not a general-purpose developer SDK. Check Google's account page and current terms before paying for access.
03 / DEEP DIVE
Compose a world prompt with useful structure
Specify environment, visual tone, landmarks and how the character should move. A focused prompt like a narrow canyon with a river and a visible mountain gives the model more navigational structure than a long adjective list. Choose first- or third-person presentation if the interface supports it.
Google's prompt tips explain the value of defining the world and character separately. Images can help communicate the starting look; they do not guarantee precise layout persistence. Expect to iterate, and keep expectations proportional to the prototype's session restrictions.
04 / DEEP DIVE
Sketch, explore and remix
World sketching uses descriptive text and optional images to create a visual starting point. Exploration then responds to movement, while remixing can produce alternative outcomes. The interactive experience differs from viewing a fixed video or static panorama.
During exploration, evaluate whether landmarks remain recognizable and controls feel coherent. In its original launch, Google documented a 60-second limit for generated exploration and said some research features such as promptable world events were not yet included. Product updates may change these rules.
05 / DEEP DIVE
Street View grounding and safety
Google introduced Street View grounding for supported U.S. locations alongside wider eligibility. It allows real-world imagery to influence imagined starting environments, but is not a guarantee of geographically precise reconstruction.
Avoid using a generated world as a factual map for navigation or professional safety decisions. Also respect privacy, content and intellectual-property policies when uploading reference images. An experimental world should be judged as creative simulation, not an authoritative replica.
06 / PRACTICAL DECISIONS
What to do, and what to avoid
Use these paired decisions as a starting point. They are editorial recommendations, not measured results from an undisclosed test.
| DO / GOOD PRACTICE | DON'T / COMMON ERROR |
|---|---|
| Check your current country, age and plan eligibility. | Copy older U.S.-only access rules as timeless facts. |
| Describe perspective, landmark and movement clearly. | Assume vague aesthetic prompts produce stable navigation. |
| Treat Street View worlds as imaginative grounding. | Use them as exact geographic or safety maps. |
07 / HANDS-ON EXAMPLE
A concise world-sketch prompt
This example emphasizes a navigable path and clear visual identity.
Third-person explorer on a windswept desert plateau. A white observatory stands to the east, a shallow canyon runs north, and sunset casts long shadows. Character: traveler carrying a yellow backpack. Camera stays behind the traveler.Choose original characters and check the current prototype's supported duration and controls.
QUESTIONS / ANSWERS
Frequently asked questions
Is Project Genie available outside the U.S.?
Google announced expanded rollout to eligible AI Ultra $200 subscribers globally in 2026; availability and conditions may vary.
Can I export a finished Unity game from Project Genie?
Google describes a research exploration prototype, not a complete Unity project export pipeline.
VERIFY / SOURCE MATERIAL
Official and primary references
Vendor model cards and launch posts are primary sources for features, prices and release status. Vendor-run benchmarks are not independent head-to-head tests. Prices, access plans and availability can change after the editorial date.
- Google: Project Genie launch
- Google: global expansion and Street View
- Google: Project Genie prompt tips
Reviewed 23 October 2026. If this article covers a forecast, that section is explicitly analysis rather than a confirmed release announcement.