EverArt Forge MCP Server
# EverArt Forge MCP for Cline

An advanced Model Context Protocol (MCP) server for [Cline](https://github.com/cline/cline) that integrates with EverArt's AI models to generate both vector and raster images. This server provides powerful image generation capabilities with flexible storage options and format conversion.
## Features
- **Vector Graphics Generation**
- Create SVG vector graphics using Recraft-Vector model
- Automatic SVG optimization
- Perfect for logos, icons, and scalable graphics
- **Raster Image Generation**
- Support for PNG, JPEG, and WebP formats
- Multiple AI models for different styles
- High-quality image processing
- **Flexible Storage**
- Custom output paths and filenames
- Automatic directory creation
- Format validation and extension handling
- Web project integration
## Available Models
- **5000:FLUX1.1**: Standard quality, general-purpose image generation
- **9000:FLUX1.1-ultra**: Ultra high quality for detailed images
- **6000:SD3.5**: Stable Diffusion 3.5 for diverse styles
- **7000:Recraft-Real**: Photorealistic style
- **8000:Recraft-Vector**: Vector art style (SVG output)
## Installation
1. Clone the repository:
```bash
git clone https://github.com/nickbaumann98/everart-forge-mcp.git
cd everart-forge-mcp
```
2. Install dependencies:
```bash
npm install
```
3. Build the project:
```bash
npm run build
```
4. Get your EverArt API key:
- Sign up at [EverArt](https://everart.ai/)
- Navigate to your account settings
- Create or copy your API key
5. Add the server to your Cline MCP settings file:
**For VS Code Extension**:
Edit `~/Library/Application Support/Code/User/globalStorage/saoudrizwan.claude-dev/settings/cline_mcp_settings.json`:
```json
{
"mcpServers": {
"everart-forge": {
"command": "node",
"args": ["/absolute/path/to/everart-forge-mcp/build/index.js"],
"env": {
"EVERART_API_KEY": "your_api_key_here"
},
"disabled": false,
"autoApprove": []
}
}
}
```
**For Claude Desktop App**:
Edit `~/Library/Application Support/Claude/claude_desktop_config.json` (macOS) or appropriate location for your OS
6. Restart Cline to load the new MCP server
## Usage Examples
Once configured, you can use Cline to generate images with prompts like:
- "Generate a minimalist tech logo in SVG format using the Recraft-Vector model"
- "Create a photorealistic landscape image with the FLUX1.1-ultra model"
- "Make me a vector icon for my project that represents artificial intelligence"
- "Generate a professional company logo as an SVG file and save it to my desktop"
### Tool Capabilities
The server provides these tools:
#### generate_image
Generate images with extensive customization options:
```
Parameters:
- prompt (required): Text description of desired image
- model: Model ID (5000:FLUX1.1, 9000:FLUX1.1-ultra, 6000:SD3.5, 7000:Recraft-Real, 8000:Recraft-Vector)
- format: Output format (svg, png, jpg, webp)
- output_path: Custom output path for the image
- web_project_path: Path to web project root for proper asset organization
- project_type: Web project type (react, vue, html, next, etc.)
- asset_path: Subdirectory within the web project assets
- image_count: Number of images to generate (1-10)
```
Notes:
- SVG format is only available with Recraft-Vector (8000) model
- Default format is "svg" for model 8000, "png" for others
- You can specify combined model IDs (e.g., "8000:Recraft-Vector")
#### list_images
List all previously generated images stored by the server.
#### view_image
Open a specific image in the default image viewer:
```
Parameters:
- filename: Name of the image file to view
```
## Troubleshooting
- **Error: Invalid model ID**: Make sure you're using one of the supported model IDs (5000, 6000, 7000, 8000, 9000)
- **Format not compatible with model**: SVG format is only available with Recraft-Vector (8000) model
- **Image not found**: Use the list_images tool to see available images
- **API authentication failed**: Check your EverArt API key
- **Images not appearing**: Check file permissions and paths
## License
MIT License - see LICENSE file for details.
TDQS
Scored across 3 tools
The three tools have clearly distinct purposes: generate_image creates new images, list_images enumerates stored images, and view_image opens a specific stored image. There is no overlap or ambiguity between these functions, making it easy for an agent to select the correct tool for each task.
All tool names follow a consistent verb_noun pattern with snake_case: generate_image, list_images, and view_image. This predictable naming scheme enhances readability and usability, with no deviations or mixed conventions.
With only three tools, the server feels thin for an image generation and management domain. While the tools cover basic operations (create, list, view), more advanced functionality like editing, deleting, or organizing images is missing, which could limit agent workflows in practice.
The tool surface is significantly incomplete for an image management server. There are notable gaps: no ability to delete, update, or organize images (e.g., tagging, moving), and no tools for managing the generation process (e.g., canceling jobs, checking status). This will likely cause agent failures when trying to perform common image-related tasks.