Optical Context MCP

Optical Context MCP is an MCP server component designed to condense large, visually structured PDFs—such as technical manuals, scanned handbooks, or product catalogs—into a small set of dense, retrievable packed images for agent workflows. Using Mistral OCR, it extracts both text and embedded images from PDFs, recomposes them into visually grouped PNG images, and exposes these artifacts for batch retrieval by AI agents or downstream tools over the Model Context Protocol.