Multi-image workflow

Combine images without losing the important details

Upload up to eight references, describe what to keep from each one, and generate a coherent new composition.

Nano Banana image lab
Supports multiple models with automatic backend routing by selected model.

Up to 8 images. Each file ≤10MB. Mix uploads and URLs; thumbnails show at 80x80.

Default is public and will be displayed in Gallery.

Public
Credits

Each run costs 20 credits

Checking your balance...

Result & live preview
We poll the Nano Banana task automatically until it finishes.
Auto refresh

Your generated image will appear here.

Recent generations
No generations yet. Create your first image!

Direct the blend with clear roles

01

Assign each reference

State which image supplies the subject, product, background, pose, palette, or typography.

02

Lock critical details

Name the facial features, logo geometry, packaging, color, and materials that must not change.

03

Describe the final layout

Specify placement, scale, camera angle, lighting direction, and the desired visual hierarchy.

Frequently asked questions

How many images can I combine?

The current generator accepts up to eight uploaded images or image URLs in one request.

Why do I need a prompt?

The prompt tells the model which information belongs to each reference and how those parts should interact.

Does generation use credits?

Yes. The page is free to open, while each generation follows the selected model and resolution credit cost.

Need a stronger starting instruction?

Browse tested prompts, copy the closest structure, and adapt it to your reference images.

Explore prompt templates