llama.cpp has been updated to support downloading and managing models via API. This enables full model lifecycle control through API calls without requiring a user interface.