What is Streaming?

LLM returns results as they’re generated, rather than waiting until completion. Better UX (users see text appear gradually), and Agents can start processing earlier. But needs proper error handling for stream interruptions.