Recraft MCP Server

A Model Context Protocol (MCP) server implementation for integrating with Recraft.ai services. This server enables AI assistants to generate images through Recraft's API using the MCP framework.

🌟 Features

• Implements MCP tools for Recraft image generation services
• Provides type-safe schemas using Zod for validation
• Supports various image generation options (styles, sizes, etc.)
• Easy integration with LLM assistants that support MCP

📋 Prerequisites

• Node.js (v18 or later recommended)
• A Recraft API key from recraft.ai

🚀 Installation

1. Clone the repository:

   git clone https://github.com/yourusername/recraft-mcp-server.git
   cd recraft-mcp-server
   

2. Install dependencies:

   npm install
   

3. Copy the environment file and configure your API key:

   cp .env.example .env
   

   Then edit .env and add your Recraft API key.

⚙️ Configuration

The server uses the following environment variables:

• RECRAFT_API_URL: The Recraft API endpoint (default: https://external.api.recraft.ai)
• RECRAFT_API_KEY: Your Recraft API key

🔧 Usage

Building the Server

npm run build

Starting the Server

npm start

Or use the provided shell script:

./start-mcp.sh

Development Mode

npm run dev

Inspecting the MCP Server

The MCP SDK includes an inspector tool to test the server:

npm run inspect

🧪 Testing

Run tests:

npm test

Run tests with coverage:

npm run test:coverage

Watch mode for tests:

npm run test:watch

🔍 Validation

The project uses Zod for schema validation. See ZOD_IMPLEMENTATION.md for details on the implementation.

📝 API

This server implements the following MCP tools:

• generate_image: Generate images from text prompts
• image_to_image: Transform an existing image based on a text prompt
• inpaint_image: Edit parts of an image using a mask
• replace_background: Replace the background of an image
• vectorize_image: Convert raster images to vector format
• remove_background: Remove the background from an image
• crisp_upscale: Upscale images with enhanced detail and clarity
• creative_upscale: Upscale images with creative enhancements
• create_style: Create a new style based on reference images
• get_user_info: Retrieve information about the current user
• save_image_to_disk: Save generated images to the local filesystem

The server also responds to a special help command which provides general information about the available tools, but this is handled as a special case in the server logic rather than as a formal tool definition.

For detailed information about the available parameters and options for each tool, use the MCP inspector tool or review the tool definitions in the code.

🛠️ Development

Linting and Type Checking

# Run type checking
npm run type-check

# Run eslint
npm run lint

# Fix linting issues
npm run lint:fix

# Run both type checking and linting
npm run validate

📄 License

MIT