microsoft/markitdown

Python tool for converting files and office documents to Markdown.

What it solves

MarkItDown is a lightweight Python utility designed to convert various file formats into Markdown. This is specifically optimized for use with Large Language Models (LLMs) and text analysis pipelines, as Markdown preserves essential document structure (like headings, lists, and tables) while remaining token-efficient and natively understood by most mainstream LLMs.

How it works

The tool provides a unified interface to convert a wide range of formats including PDF, Word, PowerPoint, Excel, HTML, EPubs, and text-based formats (CSV, JSON, XML). It can operate locally using built-in converters or integrate with cloud-based services for higher-fidelity extraction:

  • Local Conversion: Uses Python libraries to extract content from common document formats.
  • Azure Integration: Integrates with Azure Document Intelligence and Azure Content Understanding for advanced layout analysis, OCR, and multimodal extraction (including audio and video).
  • LLM-powered OCR: Supports plugins (like markitdown-ocr) that use LLM Vision to extract text from images embedded in documents.
  • LLM Image Descriptions: Can use an LLM client (e.g., OpenAI) to generate descriptions for images within files.

Who it’s for

Developers building RAG (Retrieval-Augmented Generation) pipelines or text analysis tools who need to transform unstructured documents into a clean, structured text format that LLMs can process efficiently.

Highlights

  • Broad Format Support: Handles everything from Office documents and PDFs to YouTube URLs and audio files.
  • Multimodal Capabilities: Supports speech transcription for audio and image descriptions via LLMs.
  • Extensible Architecture: Allows for 3rd-party plugins to add new conversion capabilities.
  • Flexible Deployment: Available as a CLI tool, a Python API, or a Docker image.
  • Cloud-enhanced Extraction: Optional integration with Azure services for complex tables, scanned PDFs, and video analysis.

Related

  • Project
    firecrawl/anydoc

    A fast Rust library that converts various office documents (Word, PowerPoint, Excel, PDF, etc.) into clean, LLM-ready GitHub-Flavored Markdown.

  • Project
    pymupdf/PyMuPDF

    A high-performance Python library for extracting and manipulating PDF and Office documents, featuring specialized tools for converting documents into LLM-ready Markdown for RAG pipelines.

  • Project
    ibrahimqureshae/mdflux

    MDFlux is a Windows‑only, portable desktop app that converts PDFs, Office files, HTML, CSV, images, audio, and many other formats into clean, structured Markdown. It runs locally, includes OCR for scanned PDFs, offers batch processing, and optional LLM‑based cleanup. By producing plain Markdown instead of image pages, it reduces token usage 2‑6×, lowering inference cost for RAG pipelines while keeping data private.

  • Project
    zcaceres/markdownify-mcp

    An MCP server that converts PDFs, images, audio, Office documents, and web content into Markdown format for easier ingestion by AI models.

  • Project
    magicrew/doc7

    A document conversion tool that turns PDFs, Office files, and images into AI-ready Markdown using multimodal vision models, supporting local inference for privacy and cost-efficiency.