Imagesorcery Mcp

ImageSorcery MCP is a Model Context Protocol server that gives AI agents a set of image analysis and editing tools. Once connected to an MCP compatible client such as Claude, Cursor or Cline, an agent can inspect and change image files on the user's computer through plain language requests. The toolset covers object detection and finding with YOLO models from Ultralytics, rotating, cropping and resizing, overlaying logos or watermarks with transparency handling, drawing text, arrows, circles and rectangles for annotation, reading image metadata, and optical character recognition for extracting text from documents, forms and ID cards. Example prompts include removing images that contain no animals from a folder, or cropping to a detected landmark. Processing happens locally rather than in the cloud, so image data stays on the user's machine whether the connected model is a hosted one such as Claude, Gemini or ChatGPT or a local one such as Llama or Mistral. It is aimed at developers and power users building agent workflows that need to see and manipulate images. Installation requires Python 3.10 or higher, with pipx recommended, and detailed system library notes for Docker environments are given in the project documentation. The vendor does not publish pricing on the site.