AI coding · Models

Streaming

Also called: Streamed responses, Token streaming, SSE streaming

Send tokens to the UI as soon as they are generated so people can start reading before the full answer is done.

In detail

Streaming feels like watching someone type: the first token appears immediately, which feels much faster than waiting for a full blob. Chat UIs, coding assistants, and on-page generation almost always stream.

Implementation is often Server-Sent Events over HTTP, or WebSockets. The client must handle cancel, disconnects, half-finished Markdown, and stopping mid-flight.

Compared with non-streaming calls, validation is harder mid-stream. For strict JSON, parse after the stream ends, or use APIs that support streamed structured output.

Developer infoTerm ID, DOM cues, match priority

Term ID: ai-streaming
DOM selectors: No DOM cues. This concept isn't detected directly on a page.
Priority: 1 · when several match at the same level, the higher priority wins
Version: v1 · updated Oct 5, 2026