I'm developing a conlang (constructed language) tentatively named Kezhonko. It's specifically intended to be easy for any type of vocal synthesizer- concatenative, formant, neural network, etc.- to pronounce, while still not sounding too much like any one "real" language. My ultimate goal is to make "talkloid" videos in MikuMikuDance, and I want the characters to be able to speak the same language with each other at comparable levels of proficiency. Also, if they're speaking a language no one really speaks, it'll be a bit harder for the audience to pick up any imperfections or roboticness in the performances. Also also, I just really like conlangs!
As of August 4th, 2023, I have the phonology down and am still working on the lexicon. My least favorite part!