I did a live stream for a school district earlier this year and this exact gap bit us. We had a hardware caption encoder feeding CEA-608 into our SDI chain, but vMix strips the ancillary data so the captions never reached the stream. We ended up burning open captions into the program output as a stopgap, which works but burns viewer options out of the picture. With those school and government captioning deadlines looming, even basic pass-through of caption data in vMix would be a huge win.