Introducing SOOT MCP: a shared visual context layer for AI and humans. Text got giant context windows. Visual work still relies on grids that show you 8 images at a time. We rebuilt the filesystem layer for agentic visual work. SOOT ingests millions of images, vectorizes and compresses them for instant retrieval, and renders thousands on one self-organizing canvas that you can instantly understand, and your AI can read in a single call. A universal visual context layer between everything. Your generation tools, drives, websites, social accounts, and anything else. Reply "SOOT" for a free pass to try it out. 👇