Unstructured Document Processor

Example Use Cases

• 1

  Ingest and extract text and metadata from a wide variety of document formats (PDF, DOCX, images, HTML, markdown, spreadsheets) for downstream question answering or summarization.

• 2

  Provide LLMs with document context inside Claude Desktop by running a local MCP that interfaces with the Unstructured API.

• 3

  A quick integration to preprocess and partition documents so they can be searched, chunked, and used as context in agent workflows.

Description

This MCP provides unstructured document processing capabilities, converting many file types (PDF, DOCX, images, HTML, Markdown, spreadsheets, and more) into content usable by LLMs. It acts as a local MCP server that calls the Unstructured API to partition and extract document content, then surfaces that content to Claude Desktop. The README includes quick setup steps for adding the MCP to Claude Desktop and notes that an Unstructured API key is required.