What Pandoc is
Pandoc is a command-line document converter. It reads Markdown, LaTeX, Word, HTML, reStructuredText, Jupyter notebooks, and a long list of others, and writes to an equally long list — including PDF via a LaTeX engine.
The architectural detail that explains its unusual reliability is that it does not translate format to format directly. It parses input into an internal abstract document model — headings, emphasis, citations, figures as structural objects rather than markup — and then writes that model out in the target format. This is why Pandoc handles cross-references and citations sanely where naive converters mangle them, and why adding a new output format does not require rewriting the input side. It is also why the conversions that fail tend to fail on things the model cannot represent, which is a predictable and documentable boundary rather than an arbitrary one.
Why researchers use it
- One source, every deliverable — the same Markdown file becomes the PDF for submission, the
.docxyour supervisor comments on, and the HTML for the lab site. - Citations that work everywhere —
--citeprocwith a.bibfile and a CSL style produces a correctly formatted bibliography in any of thousands of journal styles. - LaTeX to Word without tears — the single most-requested conversion in academia, and Pandoc does it including equations. The full walkthrough covers the flags that matter.
- Scriptable and reproducible — it is a command, so it goes in a Makefile, a CI pipeline, or a Snakemake rule and runs identically every time.
- Free, offline, and stable — no account, no service, and a format-conversion tool that has behaved consistently for over a decade.
Where it fits in a research workflow
Pandoc is the connector between the format you want to write in and the formats other people demand. That makes it invisible infrastructure: it belongs in the build step, not in your editing session.
The pipeline it enables is the plain-text writing workflow described in writing research papers in Markdown + Pandoc: references in Zotero exported by Better BibTeX to a .bib file, prose in Markdown in VS Code or Obsidian, citations as [@key], and Pandoc producing PDF and Word on demand. If your document also contains code that must run, Quarto is Pandoc with an execution layer on top and is the better choice. If your co-authors live in Overleaf, Pandoc becomes the export step rather than the whole pipeline.
Getting started
An hour, and the first successful conversion is the moment it clicks.
- Install Pandoc, and a TeX distribution if you want PDF output — PDF goes through LaTeX, which is the one dependency people miss.
- Convert something you already have:
pandoc paper.md -o paper.docx. Seeing a real document appear is worth more than reading the manual. - Add citations:
pandoc paper.md --citeproc --bibliography=refs.bib --csl=nature.csl -o paper.pdf. Download the CSL file for your target journal from the Zotero style repository. - The step people skip: put the command in a
Makefileor a shell script and commit it. A conversion you retype from memory each time is a conversion you will eventually get wrong the week before submission.
Pandoc vs the alternatives
| Alternative | Does it better | Pick it if |
|---|---|---|
| Quarto | The same conversion plus executing code chunks | Your document contains analysis that must run |
| Overleaf | Real-time collaborative LaTeX with no local setup | Co-authors need to edit in a browser |
| TeXstudio | Writing LaTeX directly with full typographic control | You need LaTeX’s precision and always will |
| Word’s own export | Nothing, for anything structured | You have one document and no bibliography |
Quarto is built on Pandoc, so this is less a competition than a question of whether you need the execution layer.
Cost, licensing, and your data
Free and open source under GPL, maintained by John MacFarlane and a community of contributors, with a release history stretching back to 2006. Nothing to buy, no account, no telemetry.
Everything happens on your machine, so there is no governance question at all — which is a genuine and under-appreciated advantage for anyone converting an unpublished manuscript or a document containing participant quotes. Where a cloud conversion service would mean uploading the text, Pandoc means a command. The only dependency worth planning for is the LaTeX distribution required for PDF output, which is a large install and occasionally a source of version-specific errors.
The honest review
Strengths. Pandoc solves the format problem completely and permanently. Once your writing is plain text with a .bib file, the question “what format does this journal want” stops being a reformatting job and becomes a flag. The citation handling is the specific thing that elevates it from a converter to research infrastructure — thousands of journal styles, correctly applied, from one source.
Limitations. It is a command-line tool with an intimidating options surface, and the difference between a working invocation and a nearly working one is often a flag you did not know existed. Complex layouts do not survive conversion cleanly — a heavily formatted Word template or a LaTeX document with intricate custom macros will need manual repair. PDF output requires a LaTeX installation, which is a large dependency and the usual source of first-time failures. And the docs, while thorough, are a reference rather than a tutorial, so the learning curve is steeper than the tool’s actual difficulty warrants.
Verdict. Adopt it if you write in plain text at all, or if you have ever had to convert a document between formats under time pressure. Skip it if your entire workflow is Word and always will be — the benefit accrues to people whose source is structured text. The condition that flips the answer is whether your document has code in it: if it does, use Quarto, which is Pandoc plus execution.
.docx actually looks like well before the deadline, not on the evening of it.Common questions
Is Pandoc free?
Yes — free and open source under the GPL, with no account and no paid tier. You may also need a TeX distribution for PDF output, which is likewise free.
Can Pandoc convert LaTeX to Word with equations?
Yes, and it is the standard answer to that problem. Equations convert to native Word equation objects with the right flags. Complex custom macros and intricate layouts need manual checking — the detailed walkthrough covers what survives and what does not.
Pandoc or Quarto?
Quarto if your document contains code that should execute during rendering — it is built on Pandoc and adds that layer. Pandoc if you just need format conversion for prose, or if you want the lightest possible tool in a scripted build.