Imagician

by flowy11

358 downloads
Not rated
GitHub

About

A MCP server for comprehensive image editing operations including resizing, format conversion, cropping, compression, and more based on sharp.

Details

Author
flowy11
Downloads
358
Categories
Other

- Resize images with fit and aspect ratio options
- Convert between JPEG, PNG, WebP, and AVIF formats
- Crop a specific region from an image
- Compress images with quality control
- Rotate images by any angle and flip in any direction
- Extract image metadata and batch process multiple sizes

Setting up with Highlight

This MCP is not yet compatible with Highlight’s one-click setup. However, you can still use it with Highlight by following these steps:

  1. Download and install Highlight from highlightai.com/download
  2. Navigate to the plugins tab and select "Add Custom Plugin"
  3. Configure the plugin with the settings below
    Plugin Name Imagician
    Command (node, npx, python, etc.)

    Please refer to the README for specific instructions on how to obtain API keys or other required environment variables.

  4. Enable "Start Automatically" if you want the plugin to start when Highlight launches

From the repository

Install the package globally via npm (npm install -g @flowy11/imagician) or build from source. Then add an entry to your Claude Desktop configuration file, pointing to the Imagician command. Users can then invoke tools like resize_image, convert_format, crop_image, etc. by describing operations in plain language.

resize_image

Resize an image to specified dimensions

convert_format

Convert image between formats (jpeg, png, webp, avif)

crop_image

Crop an image to specified region

compress_image

Compress an image with quality settings

rotate_image

Rotate an image by specified degrees

flip_image

Flip an image horizontally or vertically

get_image_info

Get metadata and information about an image

batch_resize

Generate multiple sizes from one image

Claude Desktop / Cursor

Paste into your MCP client config file to install this server.

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

McpServers

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

Imagician - MCP Image Editor

A Model Context Protocol (MCP) server for comprehensive image editing operations including resizing, format conversion, cropping, compression, and more.

Features

- Resize: Change image dimensions with various fit options
- Format Conversion: Convert between JPEG, PNG, WebP, and AVIF
- Crop: Extract specific regions from images
- Compress: Reduce file size with quality control
- Rotate: Rotate images by any angle
- Flip: Mirror images horizontally or vertically
- Batch Processing: Generate multiple sizes from a single image
- Metadata: Extract image information and properties

Installation

From npm

npm install -g @flowy11/imagician

From source

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

Usage with Claude Desktop

Add to your Claude Desktop configuration (~/Library/Application Support/Claude/claude_desktop_config.json):

Quick setup with npx (recommended):

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

If installed globally via npm:

{
  "mcpServers": {
    "imagician": {
      "command": "imagician"
    }
  }
}

If installed from source:

{
  "mcpServers": {
    "imagician": {
      "command": "node",
      "args": ["/path/to/imagician/dist/index.js"]
    }
  }
}

Available Tools

resize_image

Resize an image to specified dimensions with various fit options.
Parameters:
- inputPath: Path to input image
- outputPath: Path to save resized image
- width: Target width in pixels (optional)
- height: Target height in pixels (optional)
- fit: How to resize (cover, contain, fill, inside, outside)
- preserveAspectRatio: Maintain aspect ratio (default: true)

convert_format

Convert between image formats with quality control.
Parameters:
- inputPath: Path to input image
- outputPath: Path to save converted image
- format: Target format (jpeg, png, webp, avif)
- quality: Quality for lossy formats (1-100, default: 80)

crop_image

Extract a specific region from an image.
Parameters:
- inputPath: Path to input image
- outputPath: Path to save cropped image
- left: Left offset in pixels
- top: Top offset in pixels
- width: Width of crop area
- height: Height of crop area

compress_image

Reduce image file size with quality settings.
Parameters:
- inputPath: Path to input image
- outputPath: Path to save compressed image
- quality: Compression quality (1-100, default: 80)
- progressive: Use progressive encoding for JPEG (default: true)

rotate_image

Rotate an image by any angle.
Parameters:
- inputPath: Path to input image
- outputPath: Path to save rotated image
- angle: Rotation angle in degrees (positive = clockwise)
- background: Background color for exposed areas (default: #000000)

flip_image

Mirror an image horizontally or vertically.
Parameters:
- inputPath: Path to input image
- outputPath: Path to save flipped image
- direction: Flip direction (horizontal, vertical, both)

get_image_info

Extract metadata and information about an image.
Parameters:
- inputPath: Path to image file

Returns: format, dimensions, color space, file size, etc.

batch_resize

Generate multiple sizes from one image (perfect for responsive images).
Parameters:
- inputPath: Path to input image
- outputDir: Directory to save resized images
- sizes: Array of {width, height?, suffix}
- format: Output format for all sizes (optional)

Example Usage in Claude

"Please resize image.jpg to 800x600 pixels"
"Convert photo.png to WebP format with 90% quality"
"Crop avatar.jpg starting at (100, 100) with 200x200 size"
"Generate thumbnail sizes: 150px, 300px, and 600px wide"

Development

```bash

No reviews yet — be the first

Sign in to leave a review

Use Google, GitHub, or an email account so ratings stay tied to real people.

Email sign in

No reviews posted yet.