Model Context Protocol (MCP) Server

Integrate Kettio directly into your AI workflows. Our MCP server allows AI agents (like Claude Desktop) to generate, edit, and manage brand assets programmatically.

How It Works

┌─────────────────┐
│   Developer     │
│  (Claude/CLI)   │
└────────┬────────┘
         │ "Generate hero image"
         ▼
┌─────────────────┐
│   MCP Server    │
│  (This package) │
└────────┬────────┘
         │ 1. Analyze local assets
         │ 2. Call Kettio API
         ▼
┌─────────────────┐
│   Kettio API    │
│  (Your backend) │
└────────┬────────┘
         │ 3. Generate with AI models
         │ 4. Return signed URLs
         ▼
┌─────────────────┐
│   MCP Server    │
│  Downloads      │
└────────┬────────┘
         │ 5. Save to repo
         ▼
┌─────────────────┐
│  Your Repo      │
│  public/images/ │
└─────────────────┘

Installation

Option A: Use with Claude Desktop

Add to your Claude Desktop config (~/Library/Application Support/Claude/claude_desktop_config.json on macOS):

{
  "mcpServers": {
    "kettio": {
      "command": "node",
      "args": ["/path/to/AssetGenerator/mcp-server/dist/index.js"],
      "env": {
        "KETTIO_API_URL": "https://kettio.com",
        "KETTIO_API_KEY": "your-api-key-here"
      }
    }
  }
}

Option B: Use with Any MCP Client

# Install dependencies

cd mcp-server

npm install

npm run build

# Run the server

KETTIO_API_KEY=your-key node dist/index.js

Configuration

You need to configure the server with your API key.

Get Your API Key

You can generate an API key using the generator at the top of this page, or in your Profile Settings.

Available Tools

list_assets

Lists all image assets in your repository. Automatically finds common asset directories like public/images.

analyze_brand

Scans your repo for brand guideline docs and color tokens. Returns a summary of candidate files and extracted colors that the AI can use as brand context.

generate_assets

Generate new brand assets based on a text prompt. Supports brand context injection.

Parameters

prompt (string): Description of the image
count (number): Number of images (default: 3)
outputPath (string): Where to save (auto-detects if empty)
format (string): 'png', 'svg', 'webp'
brandContext (string): Additional brand guidelines
referenceAssets (string[]): Paths to style reference images

edit_assets

Edit existing assets using AI instructions. Perfect for refining generated images.

Parameters

assetPaths (string[]): Paths to assets to edit
prompt (string): Edit instructions
model (string): 'nano-banana' | 'nano-banana-pro' | 'chatgpt-image-1' | 'chatgpt-image-1-mini'

remove_bg_and_vectorize

Remove background and convert to SVG. Ideal for logos and icons.

Parameters

assetPath (string): Path to image
outputPath (string): Optional save path

open_assets

Open image files in your system's default viewer for quick preview.