[dragonwriter]

It needs a llama.cpp fork, too; so the stock runtime (based on stock llama.cpp) used by LM Studio presumably won't work for it.