ImageSorcery MCP
Local MCP server that gives AI assistants object detection, OCR, cropping, background removal and drawing tools, running computer vision models on your own machine.
✨ Try AgentScout's free AI image generator →Good for: connecting tools with MCP, building MCP servers, images and design
Use cases
About ImageSorcery MCP
ImageSorcery MCP is listed in our vision category and is aimed at builders. Local MCP server that gives AI assistants object detection, OCR, cropping, background removal and drawing tools, running computer vision models on your own machine. Pricing, as the maker states it: Free, open source (MIT). Maturity: Active. We last checked the official link on 2026-10-17.
ImageSorcery MCP alternatives
Other vision tools in our directory. See all vision tools →
Roboflow Agent
Computer-vision agent that picks models and builds/deploys vision apps; plugs into Claude Code, Codex, Cursor via MCP.
Viso Now
Describe what a camera should watch for in plain language and it builds a live vision app with detection logic, alerts and integrations, no datasets or training needed.
Ultralytics Agents
Visual canvas in Ultralytics Platform that chains YOLO models, vision-language models, conditions and actions like Slack alerts, plus official agent skills that let coding agents train, tune and export YOLO models.
NVIDIA TAO Agent Skills
NVIDIA's vision-model post-training toolkit, now packaged as agent skills so Claude Code or Codex can fine-tune, optimize and export vision foundation models from plain-English prompts.
NVIDIA VSS Blueprint
NVIDIA reference blueprint for video AI agents that search, summarize, answer questions and raise alerts over live or recorded video, with agent skills that assemble deployments from one prompt.
Autodistill
Roboflow's open-source library that uses large foundation models to auto-label images and then trains small, deployable detection or classification models with no manual labeling.
Badge for makers
Built ImageSorcery MCP? Add this badge to your README or site. It links back to this page. About badges