Thanks for sharing this.
In terms of UX, I like that there is a long list of output languages. For UI, as a user it'd be easier to use if the Example Text field was allowed to expand. For instance, let's say I have a block of text and I want to extract a series of specific outputs. The easiest way to specify that would be to dump a block of text containing all the outputs of interest, in all the contexts they appear, and then provide a list (or array) of output values under Expected Results. This way I'd receive a robust regex pattern, one that includes all variants I'm interested in capturing.