Token Stream
Domain: Internet & Web
🎯 Motivation
When an LLM generates tokens one at a time and sends each immediately. The user sees text appearing progressively.
📋 Concrete Example
Token stream: model generates The -> send. ’ capital’ -> send. ’ of’ -> send. ’ France’ -> send. ’ is’ -> send. ’ Paris.’ -> send. User sees the sentence building up, word by word. Feels faster.
🔗 Analogy
Token streaming is like a typist typing a letter while you watch - each letter appears as they type. You start reading immediately, before the letter is finished. Token streaming shows AI output as it is generated.