Haskell bindings for llama.cpp, a performant, C++-based inference engine for running large language models (LLMs) like LLaMA, Mistral, Qwen, and others directly on local hardware.