🪄 ImageSorcery MCP

Name: 🪄 ImageSorcery MCP
Author: sunriseapps

@sunriseapps

Repository Homepage

About 🪄 ImageSorcery MCP

An MCP server providing tools for image processing operations

Basic information

Config

Add this server to your MCP-compatible client using the configuration below.

{
  "mcpServers": {
    "imagesorcery-mcp": {
      "command": "/full/path/to/venv/bin/imagesorcery-mcp",
      "transportType": "stdio",
      "autoApprove": [
        "blur",
        "change_color",
        "crop",
        "detect",
        "draw_arrows",
        "draw_circles",
        "draw_rectangles",
        "draw_texts",
        "find",
        "get_metainfo",
        "get_models",
        "ocr",
        "resize",
        "rotate"
      ],
      "timeout": 100
    }
  }
}

Tools

No tools detected

We auto-extract tools from the README. The maintainer can list them under a ## Tools heading to populate this section.

Overview

What is 🪄 ImageSorcery MCP?

🪄 ImageSorcery MCP is a local image recognition and editing server for AI assistants, built on ComputerVision and the Model Context Protocol. It processes images entirely on-device, never sending them to any external servers, and works with Claude, Cursor, and Cline.

How to use 🪄 ImageSorcery MCP?

Install with pipx install imagesorcery-mcp, then run imagesorcery-mcp --post-install to download models and dependencies. Configure your MCP client to execute the imagesorcery-mcp command (e.g., in the mcpServers object). Once set up, prompt your AI with image tasks such as "crop my image" or "detect objects," and it will invoke the relevant tools automatically.

Key features of 🪄 ImageSorcery MCP

Crop, resize, and rotate images with precision
Remove background and add logos or watermarks
Draw text, arrows, circles, lines, and rectangles
Detect objects and find objects by text description
Extract text via OCR (Optical Character Recognition)
All processing stays local—no images sent to servers

Use cases of 🪄 ImageSorcery MCP

Automatically copy photos containing pets into a separate folder
Crop and center an object (e.g., a cat) in an image
Enumerate and describe form fields from a scanned document
Remove background from portraits or product photos
Overlay logos or watermarks onto batches of images

FAQ from 🪄 ImageSorcery MCP

What does the post-installation script do?

It creates a configuration file (config.toml), downloads default YOLO and CLIP models for detection and text-based finding, and attempts to install the clip package from GitHub.

Which system dependencies are required?

OpenCV requires ffmpeg, libsm6, libxext6, and libgl1-mesa-glx (installable via apt on Linux; typically included in non-virtual environments).

How do I install if pipx is not available?

Create a Python virtual environment (Python ≥3.10), activate it, run pip install imagesorcery-mcp, then imagesorcery-mcp --post-install. Use the full path to the executable in your MCP client configuration.