Setup is done using the following environment variables:
| Variable | Notes | Default |
|---|---|---|
| LATEXRENDER_OUTPUT_DIR | Directory where image files will be saved and served from | /tmp/latexrender/ |
| LATEXRENDER_TEMPLATE | Jinja2 template to use for rendering the LaTeX snippet. | template.tex in installation directory |
| LATEXRENDER_XELATEX | Location of xelatex executable. | (Auto) |
| LATEXRENDER_DVIPNG | Location of dvipng executable. | (Auto) |
| USE_X_SENDFILE | Whether to use X-Sendfile or serve the generated image directly. | True |
| LATEXRENDER_SENDFILE_ROOT | A root directory to prepended to the X-Sendfile header sent to the upstream proxy. | Empty string |
If not specified your path will be inspected to find latex and dvipng.
The default template, in Jinja2 format, is
{# vim: set ft=jinja ts=4 sw=4 tw=0 et :#}
\documentclass[12pt]{article}
\usepackage[utf8]{inputenc}
\usepackage{amsmath}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{color}
\usepackage{cancel}
\pagestyle{empty}
\begin{document}
\huge
{{ latex }}
\end{document}
Run using your favourite WSGI server:
chausette latexrender.app
gunicorn latexrender:app
It is recommended that you run this behind a frontend proxy such as nginx and use X-Sendfile to have that serve the image files.
Pass base64 encoded LaTeX snippets as the URL in one of the following forms:
http://localhost:8080/<b64latex>/
http://localhost:8080/<b64latex>.png
In the default template the surrounding math environment is not specified so you can, for example, use input of the form:
$$ e^{i\pi} + 1 = 0 $$
or:
\begin{equation*}
e^{i\pi} + 1 = 0
\end{equation}