MarkItDown — Convert documents to Markdown for LLMs

Overview

MarkItDown is a Python library and command-line tool from Microsoft that turns office documents, PDFs, web pages and archives into Markdown. It keeps the structure a model needs (headings, lists, tables, links) and drops visual layout. Conversion runs locally; cloud services (Azure, an LLM for OCR) are opt-in.

Instructions

Installation

python -m venv .venv
source .venv/bin/activate
pip install 'markitdown[all]'
markitdown --version

[all] pulls every converter. Install fewer extras to keep the environment small, or run the tool once through uv without installing it: