§ feed · storyline

Speeding up agentic workflows with WebSockets in the Responses API

OpenAI details how WebSockets and connection-scoped caching in the Responses API reduced overhead and improved latency for the Codex agent loop.

Apr 22 · 12:00:00 · primary fetch1 sourceupdated Apr 22 · 12:00:00

A deep dive into the Codex agent loop, showing how WebSockets and connection-scoped caching reduced API overhead and improved model latency.

§ sources1 publication · timeline below