I have solved this question by myself.
Finally I found a great post with interesting topic.
I added KaTeX (for mathematical formula) with a plugin to parse many things on my nodebb forum (posts content, topics title, shoutbox messages...). It runs, as necessary, before any markdown conversion (using action:markdown.updateParserRules to update parser to use MarkdownItTexmath).
I have a problem to implement this also for the "About me" part in profile. It is not parsed like posts etc.
I tried to use the "filter:parse.aboutme" hook but it's not working. Here is the reason : my parse function converts LaTeX formulas into an html specific code and it must do it before the markdown filter:parse.aboutme call. But, when this last one runs, it breaks all. For example,
<span> is changed into
Is there a way to parse the aboutme content before markdown and without losing the html code ?
Well, I found a solution using utils.decodeHTMLEntities for chars between dollars tags of KaTeX, before the markdown conversion.
I hope it's not too much insecure...