Streaming

LLM APIs can send tokens one by one as they're generated instead of waiting for the full response. What web standard enables this real-time token delivery? Flag format: CONGRESS{[standard]} Example: CONGRESS{websockets}