LaTeXML Overview

LaTeXML is a project developed at NIST that converts LaTeX files into accessible HTML documents. It is quite advanced and is used by arXiv.org for their experimental automatic conversion of preprints to HTML.

Project links:

• Official LaTeXML page: https://math.nist.gov/~BMiller/LaTeXML/
• Manual: https://math.nist.gov/~BMiller/LaTeXML/manual/

Conversion examples:

• NIST Digital Library of Mathematical Functions
• Go to https://arXiv.org, open any recent arXiv preprint and click the html link on the abstract page.

Using LaTeXML

Basic Command-Line Workflow

If you are comfortable with the command line, the workflow typically involves two steps:

latexml --dest=yourfile.xml yourfile.tex
latexmlpost --dest=yourfile.html yourfile.xml

Here is an example.

Single-Step Conversion

latexmlc --dest=yourfile.html yourfile.tex

Math Rendering Options

LaTeXML converts mathematics to MathML. Depending on browser support, this may render natively. Alternatively, you can instruct LaTeXML to use MathJax for more consistent presentation across browsers:

latexmlpost --format=html5 \
  --javascript='https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/MathJax.js?config=MML_CHTML' \
  --dest=yourfile.html yourfile.xml

Here is example using MathJax.

Installation

Installation is generally straightforward. See the official instructions at:
https://math.nist.gov/~BMiller/LaTeXML/get.html

macOS Note

While this is not listed officially, you cal also install it using homebrew.

brew install latexml

Additional Components

For full functionality (for example, to support graphics conversion), you may also need:

• ImageMagick
• A Perl interface to ImageMagick

Setting these up can be a bit involved. A simpler alternative is to edit the resulting HTML directly and insert images manually afterward.

Limitations and best practices

LaTeXML supports many standard LaTeX classes and packages (see the list at included bindings).

Some classes, such as exam.cls, are not officially supported. However, it is possible to create bindings for unsupported classes, even with the help of AI tools. (One can even use AI to create bindings: this working—though not perfect—binding for exam.cls was created with claude.ai.)

Also have a look at arXiv.org's LaTeX markup best practices for HTML conversion using LaTeXML:
https://info.arxiv.org/help/submit_latex_best_practices.html.