Tools Reference

generate_image

Create a new image from a text prompt.

generate_image({
  prompt: "A futuristic city at night with neon lights",
  model?: "gemini-nano-banana-2.1", // optional, default
  aspectRatio?: "16:9",
  imageSize?: "2K",
  thinkingLevel?: "high"
})

Parameter	Required	Description
prompt	Yes	Text description of the image to create
model	No	Model to use, see Image options
aspectRatio	No	Output shape, see Image options
imageSize	No	Output resolution, see Image options
useGoogleSearch	No	Google Search grounding (Pro only), see Image options
thinkingLevel	No	Reasoning effort, see Image options

edit_image

Edit a specific image file, optionally using reference images.

edit_image({
  imagePath: "/path/to/image.png",
  prompt: "Add a rainbow in the sky",
  referenceImages?: ["/path/to/reference.jpg"],
  model?: "gemini-2.5-flash-image"
})

Parameter	Required	Description
imagePath	Yes	Full file path to the image to edit
prompt	Yes	Text describing the modifications
referenceImages	No	Array of file paths for style transfer / guidance
model	No	Model to use, see Image options
aspectRatio	No	Output shape, see Image options
imageSize	No	Output resolution, see Image options
useGoogleSearch	No	Google Search grounding (Pro only), see Image options
thinkingLevel	No	Reasoning effort, see Image options

continue_editing

Continue editing the last generated/edited image in the session.

continue_editing({
  prompt: "Make it more colorful",
  referenceImages?: ["/path/to/style.jpg"],
  model?: "gemini-2.5-flash-image"
})

Parameter	Required	Description
prompt	Yes	Text describing the modifications
referenceImages	No	Array of reference image file paths
model	No	Model to use, see Image options
aspectRatio	No	Output shape, see Image options
imageSize	No	Output resolution, see Image options
useGoogleSearch	No	Google Search grounding (Pro only), see Image options
thinkingLevel	No	Reasoning effort, see Image options

Image options

Shared by generate_image, edit_image and continue_editing. All are optional; when one is omitted the server sends nothing for it and the model's own default applies. Not every model supports every value: Google's API rejects an unsupported combination with an explicit error, which the tool returns as-is.

Option	Values	Notes
model	gemini-nano-banana-2.1 (default), gemini-3-pro-image, gemini-3.1-flash-image, gemini-3.1-flash-lite-image, gemini-2.5-flash-image	See Model Selection.
aspectRatio	1:1, 2:3, 3:2, 3:4, 4:3, 4:5, 5:4, 9:16, 16:9, 21:9, and the panoramic 1:4, 4:1, 1:8, 8:1	Omitted, the model picks the shape: in testing every model returned landscape output for new images (about 16:9, e.g. 1376×768), while edits kept the source image's shape. Pass "1:1" explicitly for a square image.
imageSize	512px, 1K, 2K, 4K	Model default is 1K. Nano Banana 2 Lite produces 1K only.
useGoogleSearch	true / false	Grounds the image in live Google Search results (current dates, weather, events). Only gemini-3-pro-image is allowed; any other model is rejected before the API is called.
thinkingLevel	minimal, medium, high	How much the model reasons before drawing. high helps complex prompts (infographics, text in images, several consistent characters) but is slower; minimal is fastest and cheapest. Model default is medium on Nano Banana 2.1. gemini-3.1-flash-image and Lite accept only minimal / high; gemini-2.5-flash-image has no thinking levels.

get_last_image_info

Get information about the last generated/edited image.

get_last_image_info()

No parameters.

configure_gemini_token

Configure your Gemini API key (saved to ~/.nano-banana-config.json).

configure_gemini_token({
  apiKey: "your-gemini-api-key"
})

Parameter	Required	Description
apiKey	Yes	Gemini API key from Google AI Studio

get_configuration_status

Check if the API key is configured and its source.

get_configuration_status()

No parameters.