mcp -nano-banana
MCP server for Nanana AI image generation powered by Google Gemini's nano banana
claude mcp add --transport stdio nanana-app-mcp-server-nano-banana npx -y @nanana-ai/mcp-server-nano-banana \ --env NANANA_API_URL="https://nanana.app" \ --env NANANA_API_TOKEN="your-api-token-here"
How to use
This MCP server exposes Nanana AI's nano banana image generation capabilities to MCP clients (such as Claude Desktop). It supports two core tools: text_to_image for generating images from prompts, and image_to_image for transforming existing images based on a prompt. To use it, configure your MCP client to point at this server (for Claude Desktop, add an entry under mcpServers with a name like nanana, set the command to npx and the args to -y @nanana-ai/mcp-server-nano-banana, and provide your NANANA_API_TOKEN). Once running, you can request image generation by providing a text prompt to text_to_image, or supply image URLs plus a transformation prompt to image_to_image. The server uses Nanana API tokens for authentication and can optionally target a custom API URL via NANANA_API_URL if you need to point to a different endpoint.
How to install
Prerequisites:
- Node.js and npm installed on your machine
- Access to the Nanana API with a generated API token
Install globally with npm:
npm install -g @nanana-ai/mcp-server-nano-banana
Configure Claude Desktop (or your MCP client) to use the server:
- Create or edit your Claude Desktop MCP config (example path shown in README):
{
"mcpServers": {
"nanana": {
"command": "npx",
"args": ["-y", "@nanana-ai/mcp-server-nano-banana"],
"env": {
"NANANA_API_TOKEN": "your-api-token-here"
}
}
}
}
- Save the config and restart the client to load the MCP server.
Alternative local test (without global install): you can also run the server via npx directly in a project context:
npx -y @nanana-ai/mcp-server-nano-banana
Environment variables:
- NANANA_API_TOKEN: Your Nanana AI API token (required)
- NANANA_API_URL: Optional custom API URL (defaults to https://nanana.app)
Build/test tips:
- After starting, you can verify the server responds to MCP requests from your client.
- Ensure your API token has the necessary permissions for image generation.
Additional notes
Tips and common notes:
- Ensure NANANA_API_TOKEN is kept secure and not checked into source control.
- If you encounter authentication errors, re-check token validity in your Nanana account and token permissions.
- NANANA_API_URL can be useful for testing against a local or staging endpoint; omit it to use the default production URL.
- The MCP server expects mcpClients to send text_to_image and image_to_image requests with appropriate payloads as defined by the MCP protocol. Refer to the client documentation for exact request shapes.
- If you update the server version, you may need to refresh your client config in Claude Desktop to reload capabilities.
Related MCP Servers
frontmcp
TypeScript-first framework for the Model Context Protocol (MCP). You write clean, typed code; FrontMCP handles the protocol, transport, DI, session/auth, and execution flow.
shinzo-ts
TypeScript SDK for MCP server observability, built on OpenTelemetry. Gain insight into agent usage patterns, contextualize tool calls, and analyze server performance across platforms. Integrate with any OpenTelemetry ingest service including the Shinzo platform.
openai -agent-dotnet
Sample to create an AI Agent using OpenAI models with any MCP server running on Azure Container Apps
mcp-bun
Bun Javascript Runtime MCP Server for AI Agents
gemini-webapi
MCP server for Google Gemini — free image generation, editing & chat via browser cookies. No API keys needed.
lichess
A Model Context Protocol (MCP) server that enables Claude Desktop to interact with Lichess chess platform through natural language. Play games, analyze positions, manage your account, and participate in tournaments—all by simply talking to Claude.