Four open-source projects dominate document-to-Markdown conversion, and they are genuinely different tools built for different jobs. Full disclosure up front: Milldown is built on MarkItDown, so we have a horse in this race — which is exactly why this comparison names the cases where the others win.
| Engine | Made by | Built for | Reads PDF? | Standout strength | Main weakness |
|---|---|---|---|---|---|
| MarkItDown | Microsoft | LLM ingestion, breadth | Yes | Widest format support (Office, PDF, EPUB, HTML, images); fast; simple Python API | Lighter layout analysis on complex PDFs |
| Docling | IBM Research | Layout-heavy PDFs | Yes | Vision-model layout analysis: multi-column, dense tables, reading order | Slower and heavier; narrower format range |
| Marker | Community (Datalab) | Academic PDFs | Yes | Excellent on papers — equations, references, figures | PDF-centric; benefits from a GPU |
| Pandoc | Community | Publishing, any-to-any | No (output only) | Superb DOCX→Markdown; converts out of Markdown to anything | Cannot read PDFs at all; no LLM-oriented features |
Microsoft’s MarkItDown converts more formats than the other three combined — Word, Excel, PowerPoint, PDF, HTML, EPUB, images, and more — into Markdown specifically shaped for LLM consumption. It’s a small Python library with no model downloads and no GPU requirement, which makes it fast and easy to embed. The trade-off: it reads a PDF’s text layer rather than analysing its visual layout, so a gnarly multi-column report with nested tables will come out less faithfully than from Docling or Marker.
IBM’s Docling runs vision transformer models over each page to detect columns, tables, figures, and reading order. On scientific papers and financial reports with dense tables it produces measurably better structure — independent benchmarks put it ahead of MarkItDown on extraction fidelity. The cost is speed and footprint: model inference per page versus text extraction. If your corpus is layout-heavy and accuracy is worth minutes, Docling earns its keep.
Marker focuses on doing one thing extremely well: research PDFs. Equations, citations, and figure handling are its home turf, and with a GPU it’s quick. If you’re converting arXiv papers all day, start here. For mixed office documents, it’s the wrong shape.
Pandoc is the finest universal document converter ever written, and its DOCX-to-Markdown output is arguably the best there is. But the number-one misconception in this space: Pandoc cannot read PDFs. PDF is an output format for Pandoc (via LaTeX), not an input. It also has no notion of token efficiency, OCR, or LLM-oriented cleanup — it’s a publishing tool from a different era, and a brilliant one. We wrote a fuller comparison in Milldown vs pandoc.
These are libraries, though — all four assume a terminal and a Python or Haskell toolchain. If you want the MarkItDown engine with none of the setup, that’s the product we build: Milldown wraps the full engine in a native Mac app and adds what the library doesn’t have — on-device OCR for scanned PDFs, a live token counter, boilerplate cleanup, and batch watch folders. Everything runs locally.
Drag, drop, done — the full conversion engine in a native Mac app, running entirely on your machine. Free for 14 days.
Get Milldown for macOSmacOS 14+ · Apple Silicon · $39 one-time after trial (excl. VAT) · no subscription