Start from a photo you already own instead of a blank prompt.
Upload the portrait, switch to version 1 and its image-to-image mode, and describe the new setting. The person stays, the room changes. Compare faces with the original before you post it.
Attach a product, a backdrop and a prop as separate references and describe how they sit together. The image-to-image mode accepts up to ten references in one request.
Turn a phone photo into a watercolour, a clay figure or a magazine cover by describing the look. The composition comes from your photo, the style from your words.
This page holds the whole Nano Banana family. It opens on version 2, which generates from a prompt and optional references. To edit an existing picture, choose version 1 in the version bar: two modes appear, Generate and Image to Image. In Image to Image you upload one to ten photos and write what should change, because the prompt is still required. The model keeps what you did not mention and rewrites what you did, then saves the result to My Creations. There are no masks or brushes; the change is described in words. Version 1 also offers eleven aspect ratios, two more than the newer versions.
Each tab is a distinct mode or choice on this page.
Available on version 1. Upload a photo, say what to change, and the rest of the picture is kept as close to the original as the model can.
Up to Ten Photos
One subject or several to combine.
Edits in Words
Describe the change instead of painting a mask.
Edit the Edit
Feed a result back in for the next change.
Six capabilities across the family.
Change a photo by describing the edit.
Combine several photos in one request.
2, Pro, 2 Lite and 1 in one workspace.
On version 1, including 5:4 and 4:5.
On version 2 and Pro.
Every output saved for download and reuse.
Three groups, two editing workflows each.
Upload the keeper frame, switch to version 1 and describe a cleaner backdrop. Check skin, hands and edges at full size before delivery.
Give clients a painted or film-look version of a favourite shot by describing the style while keeping their photo as the source.
Attach the product photo plus a setting you like and describe the arrangement. Up to ten references fit in one Image to Image request.
Keep last year's product shot and change only the season: snow, autumn leaves or spring flowers around the same item.
Upload separate photos of relatives and describe one shared scene. Faces can shift slightly, so compare with the originals and rerun if needed.
Ask for a collectible figurine, a cartoon or a game character based on your photo, keeping the pose you uploaded.
Three actions on this page.
Select version 1 in the version bar, then the Image to Image mode.
Add one to ten photos and write what should change.
Check the edit against the original, then download from My Creations.
What it is, how editing works and what it costs.
Nano Banana is the name for Google's Gemini image models. This page gives you four versions in one workspace: version 2, Pro, 2 Lite and the original version 1.
Yes, the models come from Google. This site is an independent online studio that runs them and is not operated by Google.
Select version 1, choose Image to Image, upload your photo and describe the change. You can add up to ten images when you want to combine them.
The page opens on version 2, which generates from a prompt with optional references. The Image to Image mode appears only after you switch the version bar to version 1.
Each version supports its own set. Version 1 offers eleven ratios including 5:4 and 4:5; version 2 and Pro offer nine plus a choice of resolution.
Yes. New accounts get 10 credits once, valid for 30 days, and a version 1 edit costs 4 credits. Version 2 and 2 Lite are also open on the free tier; Pro needs a paid plan.
Dedicated pages for each next step.
Every Nano Banana version in one workspace: version 2 opens by default, and switching to version 1 adds an image-to-image mode for editing photos you already have with up to ten references.
0 / 20000
