WordPress Math Publisher

Version 3.x is required to use the LaTeX and color features as demonstrated here – see Quick Start>

WPMathPub is a WordPress plugin that embeds Pascal Brachet’s PHPMathPublisher rendering engine, allowing mathematical expressions to be written in pmath native or LaTeX then rendered and placed directly within WordPress pages, blogs, or comments as PNG images. Expressions are rendered as cache-enabled server-side PNG images using the built-in Linux GD and FreeType libraries, rather than user-dependent browser-side JavaScript MathJAX or XML-based MathML. It’s exciting to see this 30‑year‑old library gaining traction again, giving content creators more control over equation rendering through fast, cached, reusable PNG images that avoid device‑dependent display issues. See the full pmath native and LaTeX syntax here>

For the past 25 years, I’ve maintained the PHPMathPublisher rendering engine—modernizing the PHP codebase and resolving the few remaining edge‑case bugs. With the 3.x release, I’ve added size and color controls, LaTeX support, and a Gutenberg block, creating a modern, PNG‑based math‑rendering plugin for WordPress. My goal has always been to support transferable mathematical thinking through a lightweight, reliable alternative to browser‑intensive math‑publishing tools.

March 16, 2026 — WPMathPub v:3.x — Ron Fredericks, BiophysicsLab.com

Table of Contents

Mathematics is often promoted as endowing those who study it with a number of broad thinking skills … providing transferable skills which can be found across educational institutions, governments and corporations worldwide.

Cresswell C, Speelman CP (2020) https://doi.org/10.1371/journal.pone.0236153

Quick Start

Use these links to get the plugin and start publishing PNG equations on your WordPress site:

After installing the plugin, you can insert equations by placing pmath tags inside the block editor’s shortcode block, the classic editor, the interactive Gutenberg math block, or even inside standard text paragraph blocks, along with other text. Several examples are provided below. You can also optionally confirm a successful installation from the WPMathPub Admin Page>

WPMathPub Image Control and Typesetting Options

Here is an overview of the pmath controls to generate PNG images. See examples of basic and advanced native and LaTeX typesetting throughout this document.

Options to Control the pmath Shortcode Tag

[pmath size-control color-control pmath-syntax-processing]y=mx^2+b[/pmath]

Alternative option is the Math (WPMathPub) Gutenberg block. An interactive alternative to pmath tags in Shortcode. While the shortcode block has an occasional advantage as HTML code can be mixied with pmath tags. Note: pmath tags work in both the Gutenberg and Classic editors.

Optional PNG size control

Default size is 12, or site-wide selection from the tools menu.

Gutenberg block has a size slider.

Individual pmath tags can overide the default and select a custom size value:

size=x, where x in delim{[}{8,24}{]}

Optional PNG color control

Default size is “black”, or site-wide selection from the tools menu.

Gutenberg block has an interactive color picker.

Individual pmath tags can overide the default and select a custom optional color value:

color=“#RRGGBB”, a valid six-digit hex color. Or “name” where name can be :

  • black,
  • white,
  • gray/grey,
  • red,
  • green,
  • blue,
  • yellow,
  • orange,
  • purple,
  • magenta,
  • cyan,
  • teal,
  • navy 

pmath syntax processing

Default is native pmath syntax (not LaTeX)

Gutenberg block has a switch to control native pmath vs. LaTeX processing.

Individual pmath tags can overide the default native syntax processing with an optional LaTeX syntax flag: latex=1

Example 1: Simple demonstration of color, size, native vs. LaTeX parameters

Enter a shortcode block like this…

[pmath color=”#009900″ size=16]y=mx^2+2[/pmath]

[pmath color=”gray” size=20]y=mx^2+2[/pmath]

[pmath color=”red” size=14 latex=1]y=mx^2+2[/pmath]

The result looks like this …

Results

y=mx^2+2

y=mx^2+2

y=mx^2+2

Example 2: Mix WPMathPub content with text to make a statement

Enter a shortcode block like this…

[pmath size=16 color=”#9900cc”](alpha+beta)^2=alpha^2+2{alpha}{beta}+beta^2[/pmath]

Where:
[pmath size=12 olor=”black”]~alpha[/pmath] is defined as a
[pmath size=12olor=”black”]~beta[/pmath] is defined as b

The result looks like this …

Result

(alpha+beta)^2=alpha^2+2{alpha}{beta}+beta^2

Where:
~alpha is defined as a
~beta is defined as b

Example 3: Demonstrate interactive WPMathPub Gutenberg block with syntax/size/color controls and help

Create and edit symbolic equations directly within the Gutenberg block: very useful for building an equation while learning pmath native vs. LaTeX syntax. The equation is generated in real-time as you type the syntax. For example, it took me a few attempts to align b over a properly until I tried the matrix command (used to structure layout, not as a matrix equation) as shown in the equation presented here.

e^b-e^a =~ delim{[}{e^x}{]}matrix{2}{1}{b a} ~= int{a}{b}{e^x} dx

Example 4: Demonstrate a transparent PNG image on a black background (using WPMathPub Gutenberg block

This example uses a yellow equation (20 px font size) and a black background. The color was selected and copied for reuse on the “Result:” text, so that both the title and equation match the same color. This example demonstrates the true transparency of the PNG equations – showing no halo effects.

Result:

delim{|}{{1/N} sum{n=1}{N}{gamma(u_n)} - 1/{2 pi} int{0}{2 pi}{gamma(t) dt}}{|} <= epsilon/4

Example 5: Support for gradient field operators

Several physics professors have asked me to implement a gradient feature. Now it is available – yay.

The new del (aka nabla) pmath symbol opens up div, grad, and curl discussions:

  • Gradient on scalars: del f
  • Divergence on vectors: del • vec{F}
  • Curl on vectors: del x vec{F}

Include the vec pmath symbol with del to emphasize the operator’s nature as a vector:

Enter a shortcode block like this…

[pmath size=16]vec{del}f(x, ~y) ~ = ~ {partial{f}}/{partial{x}}hat{i} ~ + ~ {partial{f}}/{partial{y}} hat{j}[/pmath]

[pmath size=16]{del}f(x, ~y) ~ = ~ {partial{f}}/{partial{x}}i ~ + ~ {partial{f}}/{partial{y}} j[/pmath]

The result looks like this …

Result

vec{del}f(x, ~y) ~ = ~ {partial{f}}/{partial{x}}hat{i} ~ + ~ {partial{f}}/{partial{y}} hat{j}

{del}f(x, ~y) ~ = ~ {partial{f}}/{partial{x}}i ~ + ~ {partial{f}}/{partial{y}} j

Advanced Examples

Each of the examples below includes:

  • An overview of the pmath symbols available with this shortcode,
  • The final output is server-side PNG images,
  • A link to the raw shortcode text. You can use these shortcode text examples as starter templates for your own math publishing needs.

Advanced Example 1: Show three polynomials grouped as a matrix and delimited with only a left-hand brace.

Show three polynomials grouped as a matrix and delimited with only a left-hand brace.

Enter a shortcode block like this…

[pmath size=24/]delim{lbrace}{matrix{3}{1}{{3x-5y+z=0} {sqrt{2}x-7y+8z=0} {x-8y+9z=0}}}{~}[/pmath]

The result looks like this …

Result:

delim{lbrace}{matrix{3}{1}{{3x-5y+z=0} {sqrt{2}x-7y+8z=0} {x-8y+9z=0}}}{~}

Note: the use of delim to ensure the left brace covers the full height of the matrix. In this case, there is no corresponding right brace, so the equation ends with an empty delimiter coded to {~ }, and to ensure typography is maintained. You will know when the delimiter braces don’t match when you see unusual equation results.

Advanced Example 2: Demonstrate summation, integral, subscripts, and fractions bounded by absolute value delimiters.

Enter a shortcode block like this…

[pmath color=”blue”]delim{|}{{1/N} sum{n=1}{N}{gamma(u_n)} – 1/{2 pi} int{0}{2 pi}{gamma(t) dt}}{|} <= epsilon/3[/pmath]

The result looks like this …

Result

delim{|}{{1/N} sum{n=1}{N}{gamma(u_n)} - 1/{2 pi} int{0}{2 pi}{gamma(t) dt}}{|} le epsilon/3

Advanced Example 3: Demonstrate summation with sine and cosine trig functions.

Enter a shortcode block like this…

[pmath size=8]S(f)(t)=a_{0}+sum{n=1}{+infty}{a_{n} cos(n omega t)+b_{n} sin(n omega t)} [/pmath]

The result looks like this …

Result:

S(f)(t)=a_{0}+sum{n=1}{+infty}{a_{n} cos(n omega t)+b_{n} sin(n omega t)}

In this example, I purposely use a size=8 parameter. -the smallest usable size (and also the smallest size allowed during internal code checks) that is still readable.

Advanced Example 4: Demonstrate use of pi, infinity, integral, exponent, subscript, fraction, absolute value, matrix, delimiter, size control, and integration of text with math code.

Enter a shortcode block like this…

A formula : [pmath]delim{|}{{1/N} sum{n=1}{N}{gamma(u_n)} – 1/{2 pi} int{0}{2 pi}{gamma(t) dt}}{|} <= varepsilon/3[/pmath]

About pi :

  • [pmath]pi=sum{n=0}{+infty}{{(n!)^2 2^{n+1}}/{(2n+1)!}}[/pmath]
  • [pmath]pi=22/7-int{0}{1}{{x^4 (1-x)^4}/{1+x^2}dx}[/pmath]
  • [pmath]pi=3/4 sqrt{3}+24 int{0}{1/4}{sqrt{x-x^2}dx}[/pmath]

[pmath]delim{lbrace}{matrix{3}{1}{{3x-5y+z=0} {sqrt{2}x-7y+8z=0} {x-8y+9z=0}}}{~}[/pmath]  [pmath]delim{|}{matrix{3}{3}{a_{1 1} cdots a_{1 n} vdots ddots vdots a_{n 1} cdots a_{n n}}}{|}[/pmath]

The result looks like this …

Result:

A formula : delim{|}{{1/N} sum{n=1}{N}{gamma(u_n)} - 1/{2 pi} int{0}{2 pi}{gamma(t) dt}}{|} le varepsilon/3

About pi :

  • pi=sum{n=0}{+infty}{{(n!)^2 2^{n+1}}/{(2n+1)!}}
  • pi=22/7-int{0}{1}{{x^4 (1-x)^4}/{1+x^2}dx}
  • pi=3/4 sqrt{3}+24 int{0}{1/4}{sqrt{x-x^2}dx}

delim{lbrace}{matrix{3}{1}{{3x-5y+z=0} {sqrt{2}x-7y+8z=0} {x-8y+9z=0}}}{~}  delim{|}{matrix{3}{3}{a_{1 1} cdots a_{1 n} vdots ddots vdots a_{n 1} cdots a_{n n}}}{|}

A note about HTML display in the upper box where &lt;= is used to display <= in the documentation box above, but uses pmath syntax le in the equation itself. A maddening detail, but there are very few of these as outlined in the section below>

Advanced Example 5: Demonstrate a solution to a word problem using the Taylor series with an integral remainder and the Fourier series.

Suppose we wish to create a precise definition for a series that reads something like this:

If the sequence U = (U(n)) has terms that get arbitrarily close to l and stay that way, then there exists a natural number p such that for all n ge p, delim{|}{U(n) - l}{|} le epsilon.

In the marvelous sentence above, I use pmath latex=1 tags along with the standard text since the original science presented to me by Copilot AI used the same LaTeX typography.

Enter a shortcode block like this…

From now on :

Definition : [pmath size=18](u_{n})_{n in bbN}[/pmath] converges to [pmath size=18]l ~ doubleleftright ~ forall epsilon{gt}0, ~ exists p in bbN [/pmath] such that [pmath size=18] forall n{ge}p delim{|}{u_{n}-l}{|}{le}epsilon[/pmath]

Remark : in the case of a vector sequence, just replace [pmath size=18]delim{|}{u_{n}-l}{|}[/pmath] by [pmath size=18]delim{vert}{u_{n}-l}{vert}[/pmath]

Taylor series with integral remainder :

If [pmath size=18]f[/pmath] is a class, [pmath size=18]~ C^{n+1}[/pmath] is a class over the interval : [pmath size=18 color=”blue”] f(x)=f(0)+xf{prime}(0)+{x^{2}}/{2!}f{prime prime}(0)+…+{x^{n}}/{n!}f^{n}(0)+int{0}{x}{{(x-t)^{n}}/{n!}f^{(n+1)}(t)dt}[/pmath]

Fourier series :

Definition : Then we call the Fourier series of [pmath size=18]f[/pmath] : [pmath size=18]S(f)(t)=a_{0}+sum{n=1}{+infty}{a_{n} cos(n omega t)+b_{n} sin(n omega t)}[/pmath] with [pmath size=18]omega={2pi}/{T}[/pmath]

The result looks like this …

Result:

From now on :

Definition : (u_{n})_{n in bbN} converges to l ~ doubleleftright ~ forall epsilon{gt}0, ~ exists p in bbN such that forall n{ge}p delim{|}{u_{n}-l}{|}{le}epsilon

Remark : in the case of a vector sequence, just replace delim{|}{u_{n}-l}{|} by delim{vert}{u_{n}-l}{vert}

Taylor series with integral remainder :

If f is a class, ~ C^{n+1} is a class over the interval : f(x)=f(0)+xf{prime}(0)+{x^{2}}/{2!}f{prime prime}(0)+...+{x^{n}}/{n!}f^{n}(0)+int{0}{x}{{(x-t)^{n}}/{n!}f^{(n+1)}(t)dt}

Fourier series :

Definition : Then we call the Fourier series of f : S(f)(t)=a_{0}+sum{n=1}{+infty}{a_{n} cos(n omega t)+b_{n} sin(n omega t)} with omega={2pi}/{T}

Fix bug using “Greater Than” Symbol (>) in Gutenberg block editor

I found many “less than” perfect solutions to using the “greater than” symbol in shortcode blocks within the WordPress Gutenberg editor. I have chosen to fix this issue by using a new set of commands to replace >, <, >=, <=, and <>. Now users can enter these symbols using gt, lt, ge, le, and ne without any HTML conflict or potential bug in the Gutenberg block editor.

New handling of logic test symbols

x=y

x lt gt y
x ne y
x != y

x gt y

x ge y

x lt y

x le y

WPMathPub Admin Panel

Access the Admin Panel from Dashboard -> Tools -> WPMathPub

The admin panel has several sections useful for:
  • Side-wide PNG default settings: size and color,
  • Installation status: debugging tool,
  • PNG cache control: path, clear images, and total images.
  • Plugin uninstall policy: Keep or remove cache directory and WPMathPub metadata and options.
Screenshot

Engineering Overview

Code Architecture

WPMathPub Plugin Architecture

WPMathPub is a server-side math rendering engine for WordPress

Feature Set

  • Converts math expressions → PNG images
  • Supports both:
    • Native pmath syntax
    • Translated LaTeX input
  • Cache rendered images for performance
    • Salt on MD(pmath elements)
    • Salt on WPMathPub version

Core Processing Pipeline

WordPress Post Content
        │
        ▼
[pmath] Shortcode Handler
        │
        ▼
Attribute Processing
(size, color, latex)
        │
        ▼
LaTeX Translation Layer (optional)
        │
        ▼
pmath Expression
        │
        ▼
mathpublisher Rendering Engine
        │
        ▼
GD + FreeType PNG Renderer
        │
        ▼
PNG Image Cache
        │
        ▼
 <img> returned to page
        │
        ▼
Rendered Equation Displayed in Post

LaTeX Intermediate Layer

LaTeX input
   ↓
Regex + str_replace rules
   ↓
pmath syntax

The heart of the code consists of a “wild” and Krellian IQ-enhancing regex that converts LaTeX-styled text between the shortcode pmath tags into a function call for use with the included PhpMathPublisher.php codebase.

A short segment from my wpmathpub.php PHP code shows how the math text between shortcode tags is filtered for use by PhpMathPublisher library…

// Create a WordPress math text filter suitable for PhpMathPublisher.php
function to_phpmath($content)
{
	// Add an optional font size attribute size=xx
	// Use preg_replace_callback instead of preg_replace /e in support of php 7
        $content = preg_replace_callback('#\[pmath(\s+size=|\s?)(\d*)(\])(.*?)\[/pmath\]#si', 'wpmathfilter', $content);
	return $content;
}

A PNG image is returned that replaces the original math-styled text during display. Because of the regex and the use of an image storage gallery for previously generated PNGs, the display of math equations is very fast.

A short code snippet from the PhpMathPublisher.php library shows how PNG images are inserted into WordPress pages…

function mathimage($text,$size,$pathtoimg)
{
/*
Creates the formula image (if the image is not in the cache) and returns the <img src=...></img> html code.
*/
global $dirimg;
$nameimg = md5(trim($text).$size).'.png';
$v=detectimg($nameimg);
if ($v==0)
	{
	//the image doesn't exist in the cache directory. we create it.
	$formula=new expression_math(tableau_expression(trim($text)));
	$formula->dessine($size);
	$v=1000-imagesy($formula->image)+$formula->base_verticale+3;
	//1000+baseline ($v) is recorded in the name of the image
	ImagePNG($formula->image,$dirimg."/math_".$v."_".$nameimg);
	}
$valign=$v-1000;
return '<img src="'.$pathtoimg."math_".$v."_".$nameimg.'" style="vertical-align:'.$valign.'px;'.' display: inline-block ;" alt="'.$text.'" title="'.$text.'"/>';
}

Database Architecture

WPMathPub Database Architecture

WPMathPub intentionally keeps a very clean database footprint

wp_options

  • wpmathpub_default_color
    • site default math color: hex value
  • wpmathpub_default_size
    • default font size: number from 8 to 24
  • wpmathpub_cache_last_cleared
    • cache maintenance SQL timestamp
  • wpmathpub_activated_at
    • plugin activation SQL timestamp
    • triggers dashboard display for review/donation/dismiss after 7 days
  • wpmathpub_delete_data_on_uninstall
    • uninstall cleanup policy: 0=leave data, 1=remove data

SQL

SELECT option_name, option_value
FROM wp_options
WHERE option_name LIKE 'wpmathpub%';

SQL Results (typical)

wpmathpub_activated_at 1772748831
wpmathpub_cache_last_cleared 1772945411
wpmathpub_default_color #000000
wpmathpub_default_size 12
wpmathpub_delete_data_on_uninstall 0

wp_usermeta

  • wpmathpub_notice_dismissed
    • trigger Dashboard request for review/donation/dismiss after 7 days
      • value: 0
    • conditions disabling dashboard request (leads to value set to 1)
      • user clicks on donation button
      • user clicks on review button
      • user clicks on dismiss link
    • no further request after any response by user
      • exception: uninstall cleanup policy option active > user deletes plugin (not update) > user later reinstalls plugin

SQL

SELECT user_id, meta_key, meta_value
FROM wp_usermeta
WHERE meta_key LIKE 'wpmathpub%';

SQL Results (typical)

wpmathpub_notice_dismissed 1

References

Credits

One Comment

  1. This is a math equation in a comment
    y=mx^2 + b

    Here is the shortcode that created this equation (with spaces added to the pmath shortcode so the code will not be interpreted as math:
    [ pmath size=16]y=mx^2 + b[ /pmath]

    Reply

Leave a Reply

Your email address will not be published. Required fields are marked *