Learn how to upload your first product, describe a scene, and generate your first AI visualization in minutes.
After this guide, you’ll have a basic understanding of how to use Vision to generate your own product visualizations and create images such as this:
If you don’t have any products yet, you can use this sneaker picture to follow along the guide:
1. Upload your Product
Go to the Vision dashboard and scroll down to the Creation Tools section.
Click + New Product to open the Create Element form. Give your product a name, keep Type set to Product and click Create.
Vision opens the product’s page. Scroll down to Photo Inputs and click the slot for the view you have, such as Front.
The window that opens has two tabs. Upload Photos takes a new file from your computer and Asset Library reuses one you uploaded before. An upload is assigned to the slot and saved on its own. For best results, use a high-resolution image (at least 1000px) with good lighting.
Tip: A product holds up to 6 photos, one per view. More views help Vision keep the back and sides true to the real product. Add a simple product in Vision covers every slot. How Vision works shows which views reach Nano Banana.
2. Create a new Project
Once your product is uploaded, head to the Projects menu.
You can also access it from the dashboard quick actions section by clicking on the Create New Project button:
Give your project a name and click on Create:
Scroll down to the Scene Elements section, open its Product part and tick Include Product.
Search for your product (Sneaker) and click on Add to assign it to the project:
Describe the edits you wish to apply to your product in the Prompt text area:
For example:
A hyperrealistic high quality render in a modern city. Closeup to the ground shot. Wet road after rain. The shoe is splashing on the puddle. Shoe is bent under pressure of running according to the laws of physics. Blue, Purple, Green neons and color tones. Professional advertisement photography.
The more precise your prompt is, the closer to your idea the generated images will be.
3. Generate and Refine
Scroll down and click the Run Edits button. In about 1-2 minutes, you’ll see the generated product in the selected scene. Each image appears in the Generated Assets section while it runs. Its filters (All, Running, Failed and Finished) show what is still in progress.
Next to the Run Edits button, you can see the Estimated credits usage for the single generation.
Hover over an image and click View full size to see it in full resolution, or Download to save it.
Repeat the steps above to generate more images or refine your prompt.