XMO Image MCP Server

AI-powered image generation and editing server using Gemini Image model via XMO API.

Features

• Generate Images: Create new images from text prompts
• Edit Images: Modify existing images with natural language instructions
• Reference Images: Use multiple reference images for better editing results
• Continue Editing: Iteratively improve images without specifying file paths
• Local File Output: Save generated images locally and return only short text plus structured file metadata
• Session State: Automatic tracking of last edited image for seamless workflow

Installation

npm install
npm run build

Configuration

Set your XMO API key as an environment variable:

export XMO_APIKEY="your-api-key-here"

Or provide it directly when calling the tools.

Optional Environment Variables

| Variable | Default | Description | |----------|---------|-------------| | XMO_APIKEY | - | Your XMO API key |

Usage

As MCP Server

Add to your Claude Desktop configuration (claude_desktop_config.json):

{
  "mcpServers": {
    "xmo-image": {
      "command": "node",
      "args": ["/path/to/xmo-image-mcp/dist/index.js"],
      "env": {
        "XMO_APIKEY": "your-api-key-here"
      }
    }
  }
}

Development

# Build and run
npm run dev

# Build only
npm run build

Available Tools

get_server_info

Get server information including model name and base URL.

Parameters: None

generate_image

Generate a new image from text prompt.

Parameters:

• prompt (string, required): Text description of the image to create
• api_key (string, optional): Your XMO API key

Returns:

• Short text summary in content
• Structured metadata in structuredContent with imageCount, images[].path, images[].mimeType, images[].sizeBytes, and lastImagePath

edit_image

Edit a specific existing image file.

Parameters:

• imagePath (string, required): Full file path to the image to edit
• prompt (string, required): Text describing the modifications
• api_key (string, optional): Your XMO API key
• referenceImages (array, optional): Array of file paths to reference images

Returns:

• Short text summary in content
• Structured metadata in structuredContent with saved file paths and sizes

continue_editing

Continue editing the last image generated or edited in the session.

Parameters:

• prompt (string, required): Text describing the modifications
• api_key (string, optional): Your XMO API key
• referenceImages (array, optional): Array of file paths to reference images

Returns:

• Short text summary in content
• Structured metadata in structuredContent with saved file paths and sizes

get_last_image_info

Get information about the last generated/edited image.

Parameters: None

Returns:

• Short text summary in content
• Structured metadata in structuredContent with found, path, mimeType, sizeBytes, and modifiedAt

Output Location

Generated and edited images are saved to:

• Windows: %USERPROFILE%\Documents\xmo-images\
• macOS/Linux: ./generated_imgs/ (or ~/xmo-images/ if in system directories)

File naming format: {prefix}-{timestamp}-{randomId}.png

MCP Output Design

This server does not return image bytes in MCP tool results.

Instead, image tools:

• save images to local files
• return a short text summary in content
• return machine-readable file metadata in structuredContent

This keeps MCP stdio responses small and avoids pushing large base64 payloads into model context.

Requirements

• Node.js ^18.19.0 || >=20.6.0
• XMO API key

License

MIT

Repository

https://github.com/xm-opt/xmo-image-mcp