Nano Banana MCP Server

Generate high-quality images from text descriptions using Google's Nano Banana models.

This tool creates images from natural language prompts. For best results, be descriptive about:

• Subject and composition

• Style (photorealistic, illustration, painting, etc.)

• Lighting and atmosphere

• Colors and mood

• Camera angle and lens (for photorealistic images)

Args:

• prompt (string, required): Text description of the image to generate

• model (string): Model to use. Options:

  • 'gemini-3-pro-image-preview' (Nano Banana Pro) - Best quality, 4K, text rendering

  • 'gemini-2.5-flash-image' (Nano Banana) - Fast generation Default: Nano Banana Pro

• aspect_ratio (string): Image aspect ratio. Options: 1:1, 2:3, 3:2, 3:4, 4:3, 4:5, 5:4, 9:16, 16:9, 21:9

• resolution (string): Image resolution (Pro only). Options: 1K, 2K, 4K

• use_google_search (boolean): Enable real-time information grounding (Pro only)

Returns:

• success (boolean): Whether generation succeeded

• imageData (string): Base64-encoded image data

• mimeType (string): Image MIME type (usually image/png)

• text (string): Any accompanying text from the model

• error (string): Error message if generation failed

Examples:

• "A photorealistic portrait of an astronaut on Mars at sunset"

• "Kawaii-style sticker of a happy corgi with a transparent background"

• "Minimalist logo for 'TechStart' in blue and white, modern sans-serif font"

Error Handling:

• Returns error if GEMINI_API_KEY is not set

• Returns error if API rate limit exceeded (try again later)

• Returns error if content policy violated

Edit an existing image using text prompts with Google's Nano Banana models.

Provide an image and describe your desired changes. The model will:

• Add, remove, or modify elements

• Change style, lighting, or colors

• Adjust composition

• Apply filters or effects

The model maintains the original image's style and context while applying changes.

Args:

• prompt (string, required): Description of the edit to make

• image_base64 (string, required): Base64-encoded image data (no data URI prefix)

• image_mime_type (string, required): MIME type of the image (e.g., 'image/png', 'image/jpeg')

• model (string): Model to use. Default: Nano Banana Pro

• aspect_ratio (string): Output aspect ratio. Options: 1:1, 2:3, 3:2, etc.

• resolution (string): Output resolution (Pro only). Options: 1K, 2K, 4K

Returns:

• success (boolean): Whether editing succeeded

• imageData (string): Base64-encoded edited image

• mimeType (string): Image MIME type

• text (string): Any accompanying text from the model

• error (string): Error message if editing failed

Examples:

• "Add a small wizard hat on the cat's head"

• "Change the background to a sunset beach"

• "Make this image look like a Van Gogh painting"

• "Remove the person in the background"

Error Handling:

• Returns error if image data is invalid

• Returns error if MIME type is unsupported

• Returns error if content policy violated

Compose new images using multiple reference images with Nano Banana Pro.

Use up to 14 reference images to:

• Create group compositions

• Transfer styles between images

• Maintain character consistency across scenes

• Combine objects from different images

Limits:

• Up to 6 images of objects for high-fidelity inclusion

• Up to 5 images of humans for character consistency

• Total maximum: 14 images

Args:

• prompt (string, required): Description of how to compose the images

• images (array, required): Array of image objects with:

  • base64 (string): Base64-encoded image data

  • mime_type (string): Image MIME type

• model (string): Must be Nano Banana Pro (gemini-3-pro-image-preview)

• aspect_ratio (string): Output aspect ratio

• resolution (string): Output resolution (1K, 2K, 4K)

Returns:

• success (boolean): Whether composition succeeded

• imageData (string): Base64-encoded composed image

• mimeType (string): Image MIME type

• text (string): Any accompanying text

• error (string): Error message if failed

Examples:

• "Create a group photo of these 5 people at a beach"

• "Apply the style of the first image to the subject in the second"

• "Combine these product images into a catalog layout"

List available Nano Banana image generation models and their capabilities.

Returns information about:

• Model IDs for API calls

• Model names and descriptions

• Features and capabilities

• Maximum resolution supported

Args:

• response_format (string): Output format. Options: 'markdown' (default), 'json'

Returns:

• models: Array of model information objects with id, name, description, features, maxResolution