Glad you like RapydML, I'm its creator. I've put together RapydML as a weekend project a few years ago, same time I put together the first version of RapydScript (
rapydscript.com). Both originally relied on a combination of regexes, which makes it easy to get the project out the door but becomes a maintenance nightmare as more features are added or bugs need to be fixed. RapydScript grew in popularity much faster, so I ended up refactoring it and eventually rewriting it into a full-blown compiler, capable of handling conversions more consistently, with fewer bugs. With RapydML, however, I never had time to do the same, and with modern frameworks there is now more emphasis on merging html generation into JavaScript, so the use case for RapydML has shrunk, although there are still projects doing similar things such as Jade. So to answer your question, while a cool experiment, the project is not actively maintained at this time, since without porting it to a compiler of similar structure to RapydScript, improving the project will be painful.