MCP Hub
Back to servers

nanobanana-mcp

A security-hardened MCP server for generating and editing images using Google Gemini models. It provides tools for text-to-image creation and iterative image editing with strict input validation and secure file handling.

Updated
Feb 27, 2026

nanobanana-mcp

A hardened MCP server for Gemini image generation. Fork of ConechoAI/Nano-Banana-MCP with security fixes, strict TypeScript, and model selection.

Features

  • 3 tools: generate_image, edit_image, continue_editing
  • Model selection via NANOBANANA_MODEL env var with whitelist validation
  • Security hardened: path traversal protection, file size limits, no plaintext key storage
  • Strict TypeScript: zero any types, Zod validation on all inputs

Quick Start

Claude Code

Add to ~/.claude/settings.json:

{
  "mcpServers": {
    "nanobanana": {
      "command": "npx",
      "args": ["tsx", "/path/to/nanobanana-mcp/src/index.ts"],
      "env": {
        "GEMINI_API_KEY": "your-api-key",
        "NANOBANANA_MODEL": "gemini-2.5-flash-image"
      }
    }
  }
}

Other MCP Clients

GEMINI_API_KEY=your-key npx tsx src/index.ts

The server communicates over stdio using the MCP protocol.

Tools

generate_image

Generate a new image from a text prompt.

prompt (required): Text describing the image to create (max 10,000 chars)

edit_image

Edit an existing image with a text prompt.

imagePath (required): Full file path to the image to edit
prompt (required): Text describing the modifications (max 10,000 chars)
referenceImages (optional): Array of file paths to reference images

continue_editing

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

prompt (required): Text describing changes to make (max 10,000 chars)
referenceImages (optional): Array of file paths to reference images

Configuration

All configuration is via environment variables. No config files are written to disk.

VariableRequiredDescription
GEMINI_API_KEYYesGoogle Gemini API key
NANOBANANA_GEMINI_API_KEYNoOverride for GEMINI_API_KEY (takes priority)
NANOBANANA_MODELNoModel to use (see below)

Available Models

Model IDDescription
gemini-2.5-flash-imageFast generation, good for high-volume use (default)
gemini-3-pro-image-previewPro quality, complex prompts, better text rendering
gemini-3.1-flash-image-previewLatest model, advanced features

Output

Generated images are saved to ~/nanobanana-images/ with unique filenames. The tool response includes both the file path and the image data inline.

Security

This fork addresses the following security issues from the original:

IssueFix
API key saved to disk in plaintextRemoved config file persistence entirely
configure_gemini_token tool accepts key via MCPTool removed; keys only via env vars
Path traversal in editImagevalidatePath() checks paths resolve within $HOME or $TMPDIR
No prompt length validationCapped at 10,000 chars via Zod
Hardcoded modelNANOBANANA_MODEL env var with whitelist
Silent swallowing of reference image errorsErrors now thrown and reported
Math.random() for filenamescrypto.randomUUID()
No file size limit on readsMax 20MB
Verbose errors leak internal pathsSanitized error messages
process.cwd() fallback for output dirFixed to ~/nanobanana-images/

Development

npm install
npm run typecheck   # Type check without emitting
npm run dev         # Run with tsx (hot reload)
npm run build       # Compile to dist/

Project Structure

src/
  index.ts          # MCP server entry point (3 tool handlers)
  gemini-client.ts  # Gemini API wrapper with model selection
  file-handler.ts   # Secure file I/O with path validation
  types.ts          # TypeScript interfaces and Zod schemas

License

MIT - Based on ConechoAI/Nano-Banana-MCP

Reviews

No reviews yet

Sign in to write a review