Now that I see it laid out this way, I think Pre-parsing might be the way to go for your solution.
You'd have to set up a script that interprets your text, maybe something like this:
public void Parse(STMTextContainer(x))
{
//go thru entire string
for(int i=0; i<x.text.Length; i++)
{
string replaceValue = x.text[i]; //default value
//replace specific characters with sequences
switch(x.text[i])
{
case "A": replaceValue = "<c=cyan><q=CIRCLE>"; break;
case "B": replaceValue = "<c=cyan><q=CROSS>"; break;
etc etc...
}
//remove original character
x.text = x.text.Remove(i);
//replace with sequence
x.text = x.text.Insert(i, replaceValue);
}
}
That way, when you send the text "HELLO WORLD" to a mesh, it'll just convert it to quads, itself! If you need help writing this code for this, just let me know.
Also oops, looks like that's a typo... On line 8 of STMAutoClipData.cs, I call it "Sound Clip Data" instead of auto clip data. I'll make sure this is changed in the next update! That said, you can create new Auto Clips thru the text data inspector. (The menu that shows up when you click the [T] in the top right of any super text mesh inspector. It'll be under "Inline > Sound Clips" or "Automatic > Auto Clips"