How would you "compile" the noises into being an actual voicebank?

[ProjectKoe]

I recorded all the lines in Recstar and I'm ready to test Koe's voicebank in OpenUtau when I finally figure it out (which I will try to do again) but I don't know to actually uh. Make her work. Do I just put her in a folder? Also how would you make her "character image" appear when she's the Utau being used?

 

[Dangosan]

If you don't care about things like Kana support you can call it a day with just putting the samples in a folder depending on how the samples were recorded. In the vast majority of cases, however, voicebanks require setup files to work as intended, of which the oto.ini is the most important. This setup file contains settings that define how individual samples are processed. I'm pretty sure there is a tutorial on making otos published here. You might want to modify your regional settings before attempting to create the oto though as the synthesis engine might not parse the oto correctly if the system uses the incorrect decimal symbol (issue is well-documented in the case of original UTAU; not sure about OpenUtau). You might also want to include frequency maps if you plan to distribute your voicebank.
As for the character image, you have to specify it in another setup file called character.txt.

 

[Kiyoteru]

You can configure the oto.ini using vLabeler, made by the same developer as RecStar. RecStar, vLabeler, and OpenUtau all support unicode, so you shouldn't be running into any text encoding issues.