Imagician

Imagician is a Model Context Protocol (MCP) server that hooks into AI assistants like Claude and lets you run common image operations programmatically or with natural language commands.

Features

🖼️ Resize: Change image dimensions with several fit options like cover, contain, and fill.
🔄 Format Conversion: Switch between JPEG, PNG, WebP, and AVIF.
✂️ Crop: Pull a specific section from an image.
🗜️ Compress: Lower file size with adjustable quality settings.
↪️ Rotate & Flip: Turn images by a specific angle or mirror them.
🗂️ Batch Processing: Create multiple sizes from one source image, which is great for responsive designs.
ℹ️ Metadata: Pull image properties like format, dimensions, and file size.

Use Cases

• Automating Responsive Images: The batch_resize tool is a standout. Instead of manually creating different sizes for a srcset, you can script it to generate all the required versions from a single high-res image. This has saved me a ton of time on web projects.
• AI-Powered Content Management: When working within an AI assistant like Claude, you can manage assets without breaking your flow. For instance, you can ask it to “crop avatar.jpg to 200×200” or “convert logo.png to WebP with 90% quality,” and it just gets done.
• Standardizing Asset Pipelines: If you have a system where users upload images in various formats, you can use Imagician to automatically convert them all to an optimized format like AVIF or WebP and resize them to fit standard dimensions.

How to Use It

1. Install it through npm or build it from the source.

Installation via npm:

npm install -g @flowy11/imagician

Installation from source:

git clone https://github.com/flowy11/imagician.git
cd imagician
npm install
npm run build

2. Add the MCP server to your MCP client’s configuration file.

For Claude Code (~/.config/claude/config/settings/mcp-servers.json):

{
  "imagician": {
    "command": "npx",
    "args": ["-y", "@flowy11/imagician"]
  }
}

For Cursor (~/.cursor/mcp_settings.json):

{
  "mcpServers": {
    "imagician": {
      "command": "npx",
      "args": ["-y", "@flowy11/imagician"]
    }
  }
}

For Claude Desktop (~/Library/Application Support/Claude/claude_desktop_config.json):

{
  "mcpServers": {
    "imagician": {
      "command": "npx",
      "args": ["-y", "@flowy11/imagician"]
    }
  }
}

Available Tools

• resize_image: Resizes an image. You can set width, height, and fit mode.
• convert_format: Converts format and adjusts quality for lossy types.
• crop_image: Crops an image using left, top, width, and height offsets.
• compress_image: Reduces file size with a quality setting.
• rotate_image: Rotates an image by a given angle.
• flip_image: Flips horizontally, vertically, or both.
• get_image_info: Returns image metadata.
• batch_resize: Resizes an image into multiple specified sizes.

FAQs

Q: Can Imagician handle complex edits like layer masking?
A: No, Imagician is focused on foundational image operations like resizing, cropping, and format conversion. For more advanced edits like layer manipulation, you would need a more comprehensive tool like Photoshop or GIMP.

Q: What’s the main advantage of using this over a command-line tool like ImageMagick?
A: The primary advantage is its integration with the Model Context Protocol. This allows AI assistants to use these tools programmatically through natural language, which you can’t do with a standard CLI tool. It’s built for an AI-native workflow.

Q: Is the format conversion lossless?
A: It depends on the format. When converting to JPEG, WebP, or AVIF, you can set a quality parameter (1-100), which implies lossy compression to reduce file size. Converting to PNG would be lossless.