Skip to main content
Text to Image

Nano Banana Pro

Generate and edit high-quality images using Gemini 3 Pro image model with configurable resolution and optional Google Search grounding

View details
Try it in Ampere

Inputs

Loading input fields...
Execution Steps

Loading workflow structure...

Loading curated examples...

Overview

Nano Banana Pro generates or edits images with the Gemini 3 Pro image model. Use it for high-resolution campaign concepts, ad and social visuals, product-style creative, reference-image edits, and prompts that may benefit from optional Google Search grounding.

Use cases

  • Generate polished visual directions for ads, landing pages, launch campaigns, or social posts.
  • Edit existing product, brand, or composition references by describing the desired change.
  • Create multiple aspect-ratio or resolution variants from one brief for stakeholder review.
  • Enable Google Search grounding when current public context may help shape the image prompt.

Input tips

  • Use a prompt of 3-50,000 characters with the subject, setting, style, composition, and intended use.
  • Add image_urls only when you want the AI Tool to edit or reference existing images.
  • Provide 1-14 public source image URLs; an empty image list is treated as prompt-only generation.
  • Set num_images from 1-4, choose an aspect ratio, and select 1K, 2K, or 4K resolution.
  • Leave enable_google_search off unless the prompt needs current public context.

Expected output

The AI Tool returns one or more generated or edited image URLs with optional content type, file name, file size, width, and height, plus a short description of the result and cost metadata. The shared output view renders a single image or gallery and shows description, image count, and the Nano Banana Pro model label.

Caveats

  • Source image URLs must be public and reachable; failed downloads can stop an edit run.
  • High-resolution or multi-image requests can take longer than a single 1K image.
  • Google Search grounding can help with context, but the output still needs factual and brand review.
  • If no usable image data is returned, the run fails instead of returning an empty gallery.
  • Generated text, logos, product details, faces, and image edges should be reviewed before use.

Related AI Tools