Image-to-image editing with Nano Banana 2

Transform Photos with Image to Image AI

Start with a photo instead of a blank canvas. Describe the change in plain language and Nano Banana 2 re-renders the image while preserving composition, identity and the details you want to keep.

What You Can Do

Keep the subject, change everything else

Swap a background, a season, an outfit or an entire art direction while the face and product stay recognizable.

No masks or manual selections

Describe the edit conversationally. The model understands what to change and what to preserve without lasso tools.

Style transfer on demand

Turn a portrait into anime, pixel art, a figurine or a studio headshot using the same upload.

Batch-friendly iteration

Generate, refine the prompt and regenerate until the result matches your brief — no re-uploading needed.

Editing by description, not by hand

Traditional photo editing asks you to select, mask and layer. Image-to-image asks you to describe the outcome. That shift makes complex changes — relighting a scene, moving a product into a studio, restyling a portrait — accessible without editing experience.

When to use image-to-image instead of text-to-image

Use text-to-image when you are inventing something new. Use image-to-image when you already have a photo whose composition, subject or product must stay accurate — for example an e-commerce listing, a family portrait or a brand asset.

How It Works

01

Upload a source image

A clear JPG, PNG or WEBP photo. Portraits, products and scenes all work.

02

Describe the transformation

Pick a preset style or write the change you want, for example “place this product on a white studio background”.

03

Generate and refine

Review the result, adjust your prompt and download the final image.

Frequently Asked Questions

It means using an existing image as the starting point for an AI generation. The model keeps the structure and subject of your photo while applying the style or edit described in your prompt.
Yes. Nano Banana 2 is tuned to preserve identity, so portraits remain recognizable after restyling or background changes.
A filter applies one fixed effect to every pixel. Image-to-image re-renders the scene with an understanding of objects, lighting and depth, so a new background or outfit looks physically plausible.
JPG, PNG and WEBP uploads are supported, up to 24MB per file.

Ready to Create?

Start with one prompt or one photo — no subscription, no design experience needed.

Try Image to Image

Core Tools

Related Nano Banana 2 Tools