Nano Banana Mcp

What is Nano Banana Mcp?

Nano Banana Mcp is a Model Context Protocol (MCP) server that provides AI image generation and editing capabilities using Google’s Gemini 2.5 Flash Image API. It lets you create images from text prompts, edit existing images, iterate on results, and use reference images for style transfer.

How to use Nano Banana Mcp?

Configure the server with your Gemini API key via an environment variable in your MCP client config (e.g., GEMINI_API_KEY), a system environment variable, or the built-in configure_gemini_token tool. Then invoke commands such as generate_image, edit_image, continue_editing, get_last_image_info, or get_configuration_status through your MCP client (Claude Code, Cursor, or others).

Key features of Nano Banana Mcp

• Generate new images from text descriptions
• Edit existing images with natural language prompts
• Iteratively edit the last generated/edited image
• Use multiple reference images for style guidance
• Cross-platform file handling for Windows, macOS, Linux
• Automatic image saving with organized naming

Use cases of Nano Banana Mcp

• Generate visual assets or mockups from text descriptions
• Edit and refine existing images by describing changes
• Apply style transfer by providing reference images
• Iteratively develop a design concept through successive edits

FAQ from Nano Banana Mcp

What are the dependencies or runtime requirements?

Node.js 18.0.0 or higher and a Gemini API key from Google AI Studio are required.

How do I configure my API key?

You can set it as an environment variable in your MCP client config (recommended), as a system environment variable (GEMINI_API_KEY), or use the configure_gemini_token tool to create a local .nano-banana-config.json file.

Where are generated images saved?

Images are saved to platform-specific directories: %USERPROFILE%\Documents\nano-banana-images\ on Windows, ./generated_imgs/ on macOS/Linux, or ~/nano-banana-images/ when run from system paths.

What editing features does it support?

You can edit a specific image file with edit_image, continue editing the last generated/edited image with continue_editing, and optionally supply reference images for style transfer.

Is the API key stored securely?

The recommended method keeps the key in your MCP client’s environment config, which is contained within the MCP configuration. System environment variables and a local config file are also supported, with the config file automatically ignored by Git and NPM.