Use your own OpenAI-compatible API served locally by Llama-server or Ollama in LM Studio by fetching the model list.