How to Use Project Genie with Street View

Capture a clear Street View image

The AI works best when it has clear visual cues to understand the environment. We'll start by finding a suitable scene and taking a screenshot.

1. Open Google Maps on your computer's web browser. You'll see the familiar map interface.
2. Search for a location you want to turn into a game. Once the map loads, drag the little yellow "Pegman" icon (usually found in the bottom-right corner) onto a street. This activates Street View.
3. Navigate Street View to find a scene that looks like a side-scroller game – imagine an old arcade game where your character moves left and right across the screen. Look for a perspective that is slightly from the side, showing a clear pathway, sidewalk, or road. Avoid extremely blurry images, heavy rain, or scenes packed with hundreds of moving cars.
4. Take a screenshot of this ideal scene.

Next step ↓

Find Project Genie and upload your image

Now, let's get your image into the AI's workspace. Project Genie is part of Google's experimental "Labs" or "AI Test Kitchen" initiatives, so its exact location might shift.

1. Open your web browser and search for "Google Project Genie" or "Google AI Test Kitchen". You'll see search results that should lead to the official Google Labs page for Project Genie. Look for a link directly from labs.google.
2. Navigate to the Project Genie interface. Once on the page, you'll see a main dashboard. Look for a large box or button labeled "Upload Image", "Choose File", or a drag-and-drop area. If you're asked to sign in, use your Google Account.
3. Upload your saved screenshot. Click the upload button and select the Street View image file you saved in Step 1.

What happens next: The webpage will show your image appearing on the screen, and you'll typically see a loading animation or message like "Processing image..." or "Analyzing scene...".

Next step ↓

Adjust the interactive "physics"

Once your image is loaded, Project Genie will automatically try to map out the "interactive layers". This means the AI makes its best guess about which parts of the photo are solid objects (like the road, a brick wall, or a bench) that your character can walk on or jump over, and which parts are just background (like the sky or distant trees).

1. Observe the AI's initial mapping. You'll see your image with certain areas highlighted or subtly colored. These colors represent how the AI perceives the "physics" – where the ground is, where objects are, and what's just background.
2. Use the on-screen brush tools to refine the physics. Look for tools that resemble a paintbrush, eraser, or different colored markers. These tools usually allow you to "paint" over areas to tell the AI: "This should be solid ground," or "This object should be bouncy." For a beginner, simply focusing on marking solid ground and key jumpable objects is enough.

What happens next: As you paint, the highlighted areas on your image will change color, visually confirming your adjustments. The AI will instantly update its understanding of the environment.

Next step ↓

Generate your world and play

Now comes the exciting part – turning your static image into an interactive experience!

1. Locate and click the "Generate" or "Play" button. This button is usually prominent, often labeled clearly or represented by a large play icon. What you see: The screen will usually switch to a new view, presenting your stylised Street View scene. A small, simple character (often a pixelated figure) will appear, typically at the starting edge of your photo (e.g., the far left). What happens next: The character becomes controllable. You can usually use your keyboard's arrow keys (left, right, up for jump) or an on-screen controller to move this character. Sometimes you might need to try a different image if the AI struggled to find a clear path.