Source images (up to 8)
Edit instructions and protected details
Output resolution and aspect ratio
Upload up to eight source images, describe the exact change and what must stay untouched, and review one edited artifact before it runs in Vizify.
The request, selected options, and staged images open in Vizify for review before execution.
Source images (up to 8)
Edit instructions and protected details
Output resolution and aspect ratio
One edited image artifact
Reusable edit request and settings
Nano Banana Image Editing accepts up to eight source images, useful for edits that need more than one input photo for context.
Choose 1K, 2K, or 4K output and Auto, square, landscape, or portrait framing before the edit runs.
Name identity, geometry, logos, labels, or text that must remain intact, and the request keeps them as constraints for the edit.
Selected files are staged for the same-tab Vizify handoff and uploaded durably only after sign-in.
See current Vizify variants, supported inputs, resolutions, limitations, and other compatible tools.
Describe the exact change, such as a background swap, object removal, color adjustment, or retouch, and Nano Banana Image Editing applies it to your uploaded source images.
You can upload one to eight source images in a single request; at least one is required.
No. Nano Banana Image Editing changes photos you already have. To create a new image from a text prompt instead, use the separate Nano Banana AI Image Generator page.
Choose 1K, 2K, or 4K output resolution before you run the edit.
Call out identity, geometry, logos, labels, edges, reflections, or existing text that must remain stable, and the request carries those constraints into the edit.
No. They are removable presets for supported bounded choices; the prompt can still describe other valid requirements.
The Site stages your source images in the browser for the same-tab handoff. Vizify performs the durable authenticated upload after sign-in.
Semantic editing can rebuild nearby pixels, so untouched regions are not guaranteed to stay pixel-identical to the originals.
No. The prepared request opens in Vizify, where you can review the images and instructions before sending.
Yes. The request and resulting artifact can be reviewed, revised, and reused in Vizify.
You can prepare the task on this public page; sign-in is required to upload durably, run the edit, and save the result.
Separate the requested change from protected details. Call out identity, geometry, logos, labels, edges, reflections, and shadows when they must remain stable, and use multiple source images only when the edit genuinely needs more than one photo for context.
Public model contract: gemini-3.1-flash-image. Semantic editing can rebuild nearby pixels and cannot guarantee untouched regions remain pixel-identical. The prepared request opens in Vizify first; no task runs until you review and send it.