Making LLMs Real-time – How to stream LLM responses to the client in real-time