Cheatsheets:Pandoc
Quick reference for smart people — part of our dev cheatsheets collection.
Pandoc is a universal document converter: write once in Markdown, output to HTML, PDF, DOCX, EPUB, LaTeX, or slides — and convert back from nearly any format.
New to Pandoc? Try the 20-minute tutorial or skim the manual.
Commands assume Pandoc 3.x — check yours with
pandoc --version.
Basic example
pandoc pancake.md -o pancake.html
turns
!pancake.md
# Pancakes
A quick pancake recipe for two.
into
!pancake.html
<h1 id="pancakes">Pancakes</h1>
<p>A quick pancake recipe for two.</p>
Pandoc automatically infers the file formats from file extensions!
Converting between formats
| Command | Conversion |
|---|---|
pandoc report.md -o report.html |
Markdown → HTML |
pandoc report.md -o report.pdf |
Markdown → PDF (see PDF output) |
pandoc report.md -o report.docx |
Markdown → Word |
pandoc book.md -o book.epub |
Markdown → EPUB e-book |
pandoc slides.md -t revealjs -s -o slides.html |
Markdown → HTML slide deck |
pandoc page.html -o page.md |
HTML → Markdown |
pandoc data.csv -t markdown |
CSV → Markdown table |
pandoc ch1.md ch2.md -o book.html |
several files, concatenated |
When -o is not provided, output goes to stdout
pandoc notes.docx -t markdown | less
Similarly, when no input file is specified, pandoc reads from stdin
echo "# Hi" | pandoc -t html
Since pandoc infers the file format from the file extension and not from file content, a file without extension (e.g.,
-o out) causesUnknown output formaterror.
PDF output
Pandoc needs a typesetting engine to output PDF.
The default engine is LaTeX.
pandoc report.md -o report.pdf # default engine: pdflatex
pandoc report.md -o report.pdf --pdf-engine=xelatex
pandoc report.md -o report.pdf --pdf-engine=weasyprint # HTML-based, no LaTeX needed
pandoc report.md -o report.pdf --pdf-engine=typst # pandoc >= 3.1.5
If LaTeX is not installed, you may get
pdflatex not founderror.--pdf-engine=weasyprintor--pdf-engine=typstcircumvents the problem.
Standalone documents, TOC & metadata
A bare conversion emits a fragment — no <head>, no title, no TOC. Add -s for a full document:
pandoc report.md -s -o report.html # full HTML page
pandoc report.md -s --toc -o report.html # ... with a table of contents
pandoc report.md -s --toc -N -o report.docx # TOC + numbered sections
Title, author, date must be specified in a YAML block at the top of the source Markdown file:
---
title: "Pancake report"
author: Ada
date: 2025-06-01
---
...or on the command line:
pandoc report.md -s -o report.html -M title="Pancake report"
--tochas no effect without-s— a fragment has no place to put the table of contents.Also, Pandoc reads its own pandoc markdown dialect by default. Other dialects:
-f gfm(GitHub) or-f commonmark.
Common options
| Option | Effect |
|---|---|
-f FORMAT / --from=FORMAT |
input format: markdown, gfm, commonmark, html, docx, epub, csv, ...
|
-t FORMAT / --to=FORMAT |
output format — same list, plus pdf, revealjs, beamer, typst
|
-o FILE |
write to FILE instead of stdout |
-s / --standalone |
full document, not a fragment |
--toc |
table of contents; --toc-depth=2 limits the depth
|
-N / --number-sections |
number the headings |
-M key=value |
set metadata, e.g. -M title="Report"
|
--pdf-engine=ENGINE |
pdflatex (default), xelatex, lualatex, weasyprint, typst |
--reference-doc=FILE |
style template for docx / pptx / odt |
--highlight-style=NAME |
code highlight theme: pygments (default), tango, zenburn, monochrome, ... |
--wrap=auto|none|preserve |
line wrapping in text output |
--extract-media=DIR |
where images from docx / epub are saved |
--citeproc |
process citations and bibliography |
--lua-filter=FILE |
transform the document with a Lua script |
--list-input-formats / --list-output-formats |
list every supported format |
docx → Markdown keeps the images — Pandoc extracts them into a
media/folder next to your output by default;--extract-media=DIRchanges that directory.
For more
- Pandoc manual — the official, exhaustive reference
- Pandoc tutorial — the recommended quick tour
- Try Pandoc — run conversions in the browser
- Pandoc filters — extending Pandoc with Lua scripts
- Cheatsheets:Quarto — a publishing system built on Pandoc