AI WORLDHow to Use Project Genie with Street View
Turn everyday street photos into your own playable virtual worlds with Google's clever experimental AI.
Imagine taking a quick snapshot of your local café, your childhood street, or a famous holiday spot, and instantly turning it into a video game level you can run and jump through. This guide will show you how to use Google's Project Genie to transform flat Street View images into interactive, playable virtual worlds.
Project Genie is a generative AI model (a type of artificial intelligence that can create new things, like images or interactive environments, rather than just analysing them). Think of it as a digital stage designer. It takes a single 2D image and guesses where the solid ground is, where the obstacles are, and how a virtual character should move through the space.
Best of all, you do not need to know a single line of code to make this happen. Here is how to get started.
- Get a Google Account: You will need a standard Google login to access Google’s experimental platforms.
- Find a Street View image: Open Google Maps on your phone or computer, search for a cool street, and take a clear screenshot of the scene.
Capture a clean street scene
The AI works best when it has clear visual cues to work with. Look for a Street View scene that looks like a classic side-scroller game.
Look for a perspective that is slightly from the side, showing a clear pathway, sidewalk, or road. Avoid images that are extremely blurry, taken in heavy rain, or packed with hundreds of crowded moving cars.
Upload your image to the workspace
Open the Project Genie experimental page (usually found within Google’s AI Test Kitchen or Labs portals).
- Look for the main dashboard interface. You will see an option that looks like Upload Image or a drag-and-drop box.
- Select your saved Street View screenshot and upload it.
- The AI will take a few moments to process. During this time, it is analysing the shapes, lines, and depth in your photo to understand where a character can physically walk.
Let the AI map the "physics"
Once the image is loaded, Project Genie will automatically try to map out the "interactive layers".
In simple terms, this means the AI decides which parts of the photo are solid objects (like the road, a brick wall, or a bench) and which parts are just background (like the sky or distant trees).
You can use the simple mouse brush tools on the screen to paint over any areas you want to make extra bouncy, solid, or slippery.
Press play and explore
Now comes the magic. Click the Generate or Play button on the screen.
The AI will generate a little character at the starting edge of your photo. You can now use your keyboard's arrow keys or an on-screen controller to move left, right, and jump.
As you move, the AI will dynamically imagine what lies just off the edges of your original photo, creating an endless, seamless world based on the style of your Street View snapshot.
- Using top-down photos: If you take a screenshot looking straight down at the ground from a bird's-eye view, the AI will struggle to understand where the "ground" is. Stick to eye-level street perspectives.
- Expecting Hollywood graphics: Remember, Project Genie is designed to create charming, retro-style interactive environments. It converts your photo into a stylised playground, so do not worry if some details look a bit blocky or dreamlike.
- Too much clutter: If your photo has too many overlapping power lines, crowds of people, and signs, the AI character might get stuck. Start with simple, open streets first.
Take a screenshot of your own driveway or front street right now. Head over to your Google Labs dashboard, upload the image, and spend two minutes jumping over your own mailbox in a virtual world!
✦ Original step-by-step guide by AI World Co.'s AI editorial team. Written in plain language, reviewed for accuracy.
← Back to all stories