Llama.cpp

models:
  - name: <MODEL_NAME>
    provider: llama.cpp
    model: <MODEL_ID>
    apiBase: http://localhost:8080