@ Squibcakes

what you are after looks a bit like a Multiplexer for NMEA-Data.
I have solved that problem the following way:

As there is no defined timing between the different Talkers
and I didn't want to miss sentences (i.e. two devices are talking at the same time)
I used one 12F629 per serial input line.
As the 12F doesn't have enough memory to store the full seetence I isolate the information bytes I need and skip all the rest.

An aditional PIC (12F or 16F) polls the "receivers", assembles the "combined" sentence and sends it out on one single serial line.

This solution offers high flexibility at about $2 per Talker.

rgds