Document Understanding MCP Server

A document reading, search, and metadata server to provide access to PDFs (and, in the future, other formats) to LLMs via the Model Context Protocol

Local serverstdioPython

What is the Document Understanding MCP Server?

A document reading, search, and metadata server to provide access to PDFs (and, in the future, other formats) to LLMs via the Model Context Protocol (MCP)

What the server does

MCP server providing tools to extract text, metadata, layout, and search documents (primarily PDFs). This server implements the Model Context Protocol (MCP) specification, allowing AI models to interact with PDF documents through a standardized interface.

The Document Understanding MCP Server provides a set of tools for extracting information from PDF documents, including:

More MCP options

Compare MCP servers by category, or browse hosted MCP servers if you prefer an endpoint over a local process.

Maintainer setup documentation.

Recorded tools

Check the linked publisher documentation and your connected client's tool list. Names and availability can vary with the server version and account permissions.

ToolWhat it does
get_pdf_working_directory> Check response for working_directory.

Configuration

VariableDescriptionRequired
DOCUMENT_UNDERSTANDING_BASE_PATHFilesystem location the server is allowed to use.Optional
DOCUMENT_UNDERSTANDING_SANDBOXConfiguration value read at startup.Optional
ALLOW_ANY_PATHFilesystem location the server is allowed to use.Optional
PDF_SERVER_BASE_PATHFilesystem location the server is allowed to use.Optional

Example prompts to try

  • “Use Document Understanding to Passwords.”
  • “Use Document Understanding to get pdf working directory.”

Frequently asked questions

Yes, table extraction requires a Java Runtime Environment (JRE). The OCR fallback for scanned pages requires Tesseract OCR. The server will refuse to start if either is missing, unless you use `--allow-no-java` or `--allow-no-tesseract` flags.