Speeding up agentic workflows with WebSockets in the Responses API

Practical AI: Tools, Models & Frameworksapi

What changed

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

Why it matters

A concrete addition to Practical AI: Tools, Models & Frameworks: it changes what's available to builders today rather than being general commentary.

How it compares

Related prior coverage to compare against:

  • Unrolling the Codex agent loop
  • New tools and features in the Responses API
  • Open LLM Leaderboard: DROP deep dive

Sources