It's a pure JavaScript application, so it runs entirely in the browser and doesn't require any special server-side stuff, it only needs static files. It's also a public GitHub project, so you can download a snapshot of all the files as a ZIP: https://github.com/radishengine/denormalize/archive/master.zipSorry for the delay, been changing jobs and not had a lot of time to internet! I did a manual extraction job by capturing frames from dosbox, boy I wish I'd had this strip a month ago! Thanks so much for that, there's a lot of characters I would have missed and lots from the other fonts that are hard to extract. really appreciate it!
I finally took a look at the fonts. They are MGL-compressed in the standard way, but the actual data is raw pixel data. That means there's no way to tell how big each character is supposed to be, so there's no easy way to display the characters in a sensible way.
What I did instead is import the decompressed font data into GIMP as "Raw image data" doing trial-and-error with different widths until it looked right. So, here are the Normality fonts as four (very very tall) PNG images:
https://imgur.com/a/CMPR3qZ
How are the voice files? I've always wondered if there's any un-heard speech hidden and unused in the resources.
EDIT: one other thing I'd love to see are the Save/Load, Quit To Dos and Options background images. There's usually text over them so it's hard to see the original background images. Can these be extracted by the tool from the non-map resources?