Unpacking the Codex agent loop
This article explains how the Codex CLI agent loop works under the hood. It covers the iterative cycle of receiving context, deciding on actions, executing tools, and observing results. It highlights …
Responses API
Accelerating agentic workflows using WebSockets in the Responses API
This article explains how reworking the Codex agent loop with WebSockets and connection-scoped caching reduces per-turn latency and resource overhead in agentic workflows. Persistent connections elimi…
From model to agent: Building a computer environment for the Responses API
This article explains OpenAI's agent infrastructure through three core components: the Responses API for stateful, multi-step agent workflows; the shell tool that lets agents execute commands and insp…
A developer's handbook for GPT-5.6
GPT-5.6, OpenAI's latest flagship model, is changing how startups build AI agents. By deliberately routing requests across reasoning effort levels and using the Responses API, teams cut costs and late…