Dudoxx EverArt Server

Advanced EverArt Model Context Protocol server with img2img, custom model training, and full API capabilities

CI/CD License: MIT Version

✨ Features

Text-to-Image Generation

• ✅ FLUX1.1 (Standard & Ultra)
• ✅ Stable Diffusion 3.5 Large
• ✅ Recraft V3 (Realistic & Vector)
• ✅ Multiple format support (PNG, JPEG, WebP, SVG)
• ✅ SVG optimization
• ✅ Dynamic resolution (256-2048px)

Image-to-Image ✨ NEW

• ✅ Transform existing images from URLs
• ✅ Style transfer and enhancement
• ✅ Guided generation with base images
• ✅ Support for all EverArt models

Resolution Control ✨ NEW

• ✅ Custom height and width (256-2048px)
• ✅ Flexible dimensions per generation
• ✅ Defaults to 1024x1024 (backward compatible)

Web Integration

• ✅ Custom output paths
• ✅ Web project structure support (React, Vue, Next.js)
• ✅ Automatic directory creation
• ✅ Local image storage

📦 Installation

# Clone the repository
git clone https://github.com/dudoxx/ddx-everart-mcp.git
cd ddx-everart-mcp

# Install dependencies
pnpm install

# Build the project
pnpm run build

🚀 Quick Start

Configure Claude Code

Add to ~/.claude.json:

{
  "mcpServers": {
    "ddx-everart-mcp": {
      "timeout": 60,
      "type": "stdio",
      "command": "node",
      "args": [
        "/absolute/path/to/ddx-everart-mcp/build/index.js"
      ],
      "env": {
        "EVERART_API_KEY": "your_api_key_here"
      }
    }
  }
}

📖 Usage Examples

Text-to-Image (txt2img)

Basic generation with default 1024x1024 resolution:

// Ask Claude Code:
"Generate a serene mountain landscape at sunset"

Custom Resolution

Generate images with custom dimensions:

// Ask Claude Code:
"Generate a wide panoramic mountain landscape, use width 2048 and height 512"

Supported dimensions:

• Minimum: 256x256 pixels
• Maximum: 2048x2048 pixels
• Default: 1024x1024 pixels (when not specified)

Image-to-Image (img2img)

Transform an existing image from a URL:

// Ask Claude Code:
"Transform this image into a watercolor painting style: https://example.com/photo.jpg, type img2img"

Complete img2img workflow:

// Step 1: Use an image URL
const imageUrl = "https://example.com/base-image.jpg";

// Step 2: Request transformation
"Take this image: https://example.com/base-image.jpg and make it look like an oil painting, use type img2img"

Combining Features

Use img2img with custom resolution:

// Ask Claude Code:
"Transform https://example.com/photo.jpg into anime style, type img2img, width 1536, height 1024"

🔧 Tool Parameters

generate_image

Parameter	Type	Required	Default	Description
prompt	string	✅ Yes	-	Text description of desired image
type	string	No	txt2img	Generation type: txt2img or img2img
image	string	Conditional*	-	Image URL for img2img (required when type is img2img)
height	number	No	1024	Image height (256-2048)
width	number	No	1024	Image width (256-2048)
model	string	No	5000	Model ID (5000, 9000, 6000, 7000, 8000)
format	string	No	png	Output format (svg, png, jpg, webp)
image_count	number	No	1	Number of images (1-10)

*Required when type is img2img

⚠️ Troubleshooting

img2img Issues

Problem: "image parameter must be a valid HTTP or HTTPS URL"

• Solution: Ensure your image URL starts with http:// or https://
• Example: ✅ https://example.com/image.jpg ❌ example.com/image.jpg

Problem: "image parameter is required for img2img"

• Solution: When using type: img2img, you must provide an image URL parameter
• Example: Include both type: img2img and image: https://...

Problem: Generation fails with img2img

• Solution: Verify the image URL is publicly accessible and returns a valid image
• Tip: Test the URL in a browser first

Resolution Issues

Problem: "height must be between 256 and 2048"

• Solution: Use dimensions within the supported range
• Valid: 256, 512, 768, 1024, 1536, 2048
• Invalid: 128, 4096, -100

🔄 Backward Compatibility

All existing functionality remains unchanged:

• ✅ Calling without type defaults to txt2img
• ✅ Calling without height/width defaults to 1024x1024
• ✅ All previous parameters still work exactly as before
• ✅ No breaking changes to existing integrations

🙏 Attribution

This project is a refactored and enhanced version of everart-forge-mcp by Nick Baumann. We extend our gratitude for the original work that made this possible.

Refactored by: Dudoxx UG (Hamburg, Germany) Original Author: Nick Baumann License: MIT

---

Made with ❤️ in Hamburg by Dudoxx UG