should the LLM be connected via a llama.cpp instance outside of pinokio?

I am trying to finish setup of Odysseus
should the LLM be connected via a llama.cpp instance outside of pinokio?

I tried various syntax for local LLM in various configuration places...got errors
I have tried adding llama.cpp from cookbook dependency menu....did not load/work
spoke to an external AI instance with some varying responses that seemed to recommend not using pinokio....

Your system (OS / GPU / RAM / VRAM / etc.)
Win 11, Intel cpu, 96gb RAM, Nvidia 5090 24gb VRAM, SSD