Recraft MCP Server

🚀 Reconstructing MCP Server

An implementation of a Model Context Protocol (MCP) server integrated with the Recraft.ai service. This server enables AI assistants to generate images using the MCP framework via Recraft's API.

🚀 Quick Start

This MCP server allows AI assistants to generate images through Recraft's API using the MCP framework, providing a seamless experience for image generation in AI applications.

✨ Features

• Implements the MCP tool for Recraft's image generation service.
• Utilizes Zod for type-safe schema validation.
• Supports multiple image generation options (styles, sizes, etc.).
• Facilitates integration with MCP-compatible LLM assistants.

📦 Installation

Prerequisites

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

Steps

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.

🔧 Technical Details

Configuration

The server uses the following environment variables:

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

Usage

Build the server

npm run build

Start the server

npm start

Or use the provided shell script:

./start-mcp.sh

Development mode

npm run dev

Check the MCP server

The MCP SDK includes a checking tool to test the server:

npm run inspect

Testing

Run tests:

npm test

Tests with coverage:

npm run test:coverage

Tests with watch mode:

npm run test:watch

Validation

This project uses Zod for schema validation. Refer to ZOD_IMPLEMENTATION.md for implementation details.

📚 Documentation

API

This server implements the following MCP tools:

• generate_image: Generate an image from a text prompt.
• image_to_image: Transform an existing image using a text prompt.
• inpaint_image: Edit a part of an image using a mask.
• replace_background: Replace the background of an image.
• vectorize_image: Convert a bitmap image to a vector format.
• remove_background: Remove the background from an image.
• crisp_upscale: Upscale an image with enhanced details and sharpness.
• creative_upscale: Upscale an image with creative enhancements.
• create_style: Create a new style using a reference image.
• get_user_info: Get information about the current user.
• save_image_to_disk: Save the generated image to the local file system.

The server also responds to a special help command, providing general information about the available tools. However, this is handled as a special case in the server logic, not as an official tool definition.

For detailed information on the available parameters and options for each tool, use the MCP checking tool or view the tool definitions in the code.

Development

Code checking and fixing

# 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