nano-banana-2

Gemini image generation creates visuals from descriptive text prompts by submitting your input to the gemini-3.1-flash-image-preview model and saving the resulting output files directly in your local .nano-banana/ directory.

Editing existing images applies natural language transformations to local image files, processing your requested modifications through the Gemini model and saving the updated visual assets into the project's local .nano-banana/ directory.

Grounded image generation produces visuals informed by current web and image references, using live search data to ensure the generated output reflects up-to-date styling and factual context for your creative workflows.

A GEMINI_API_KEY is required to authenticate requests and access the gemini-3.1-flash-image-preview model for text-to-image generation, image editing, and search-grounded visual creation tasks within your project directory.

Organizing generated image assets is handled automatically by storing all visual outputs under the .nano-banana/ folder in your project directory, using local directory organization for straightforward asset management and iterative refinement.

Grounded image generation requires live search references to produce images informed by current web data, ensuring that creative workflows requiring rapid ideation and up-to-date reference styling maintain factual accuracy across diverse subjects.