"author:jing-hy" matching MCP servers:

Enables text-only LLMs to understand images by converting them into structured text grids (colors, textures, regions) and OCR via MCP tools. Runs locally with zero external dependencies, providing a skill-based methodology for detailed image analysis.