Glossary · Updated Sep 2, 2026

Streaming responses

Delivering a model's output progressively as it is generated instead of waiting for the full response.

Definition

Streaming sends tokens, audio chunks, or partial results as they are produced, which reduces perceived latency and allows interfaces to show progress. It is standard in chat products and voice agents and is supported by most APIs through server-sent events or websockets.

Why it matters when choosing a tool

Streaming is a small technical detail that strongly affects how responsive a tool feels, especially for long outputs.

Where you will meet it

AI Assistants, AI Audio & Voice

Related terms

Latency · Inference · API access

Tools where this matters

Reviewed products in the related categories