nano-banana-image

To generate images from text prompts, send a generate action to the API endpoint with your authorization token, a descriptive prompt, and a selected model like nano-banana to create the visual asset.

Yes, you can edit an existing image without a separate mask by providing a source image URL along with natural-language editing instructions to apply iterative in-place modifications.

The nano-banana model provides standard results, nano-banana-2 offers improved quality, and nano-banana-pro delivers the highest detail for generating and editing creative assets.

Yes, the API workflow supports asynchronous callbacks and task polling, allowing you to manage long-running image generation or editing tasks without blocking your application.

You can control the aspect ratio and resolution of generated images by passing these optional parameters in your API request alongside the required action and model selection.

Natural-language image editing relies heavily on descriptive accuracy; ambiguous prompts may produce unexpected modifications, meaning iterative edits are required to achieve precise creative asset results.