Server facts
gemini-nanobanana-mcp
Project
- Author
- Junhan2
- Repository
- github.com/Junhan2/gemini-nanobanana-mcp
- License
- MIT
- Category
- AI Models & Agents
Tech
- Runtime / language
- node
- Transports
- stdio
- Remote available
- No
- Authentication
- Not specified
Activity
- GitHub stars
- 11
- Last commit
- Data as of
Tags
- Claude
- chatgpt
- gemini
- nanobanana
Overview
Image creation and revision sit together in Gemini Nanobanana’s toolset, which connects MCP clients to Google’s Gemini 2.5 Flash Image preview. The generate_image tool turns a written description into an image, while edit_image applies requested changes to an existing image supplied as a file path or base64 data. For workflows involving several source images, compose_images uses a guiding prompt to create a combined result. The style_transfer tool takes a base image and a separate style reference, with an optional prompt to direct the transformation.
The documented setup covers Claude Desktop, Claude Code, and Cursor, using a Google AI Studio API key. Generated images are saved automatically by default, with settings for changing the destination or disabling automatic disk saves so images remain in the conversation. Users can also request a specific output path. For Claude Code on Mac, optional hooks open generated images in Quick Look. Configuration additionally exposes logging controls, and PNG is the default output format.
gemini-nanobanana-mcp setup notes
- gemini-nanobanana-mcp records stdio only; confirm a network transport upstream before planning remote access.