This repo is a nix flake for easily running various LLMs with llama.cpp. Right now all commands run llama.cpp with a model downloaded from huggingface, but could be easily extended to support other model sources and/or alternatives to llama.cpp.