What it does
Microsoft's MarkItDown converts office documents, PDFs, HTML, and other inputs into Markdown. It provides a CLI, piped input, format-specific dependencies, and optional integrations.
Use a local file or piped input and write Markdown through the CLI or Python API. Format extras enable office/PDF/audio support, while plugins are disabled by default. Built-in extraction targets headings, lists, tables and links for analysis rather than high-fidelity page reproduction. Optional Azure Document Intelligence or Content Understanding routes send work to billable cloud APIs; some add-ons cost extra. The vision OCR plugin silently skips OCR without an LLM client. Resource access uses the process's privileges, so applications can choose narrower local-file or stream conversion methods.
Strengths
- Converts several document families into a common structured-text representation.
Watch out for
- The project prioritizes text-analysis output over high-fidelity presentation.
Best for
- Preparing mixed document collections for text analysis