# LiteLLM gateway: one OpenAI-compatible endpoint fronting local + cloud models.
# The backend references the `model_name` aliases (local-triage / cloud-reason).

model_list:
  # Local triage model served by Ollama on the host (Apple Silicon).
  - model_name: local-triage
    litellm_params:
      model: ollama_chat/qwen2.5:7b-instruct
      api_base: os.environ/OLLAMA_BASE_URL

  # Google Gemini (Google AI Studio free tier eligible).
  - model_name: gemini-economy
    litellm_params:
      model: gemini/gemini-2.5-flash
      api_key: os.environ/GEMINI_API_KEY

  - model_name: gemini-premium
    litellm_params:
      model: gemini/gemini-2.5-pro
      api_key: os.environ/GEMINI_API_KEY

  # Cloud reasoning model (deep troubleshooting) via OpenRouter.
  - model_name: cloud-reason
    litellm_params:
      model: openrouter/anthropic/claude-3.7-sonnet
      api_key: os.environ/OPENROUTER_API_KEY

  # Direct DeepSeek option.
  - model_name: deepseek-reason
    litellm_params:
      model: deepseek/deepseek-chat
      api_key: os.environ/DEEPSEEK_API_KEY

  # DeepSeek via OpenRouter (alternative routing).
  - model_name: openrouter-deepseek
    litellm_params:
      model: openrouter/deepseek/deepseek-chat
      api_key: os.environ/OPENROUTER_API_KEY

general_settings:
  master_key: os.environ/LITELLM_MASTER_KEY

litellm_settings:
  drop_params: true
  request_timeout: 120