OpenAI details Agents API: hosted agent loop, developer-controlled code execution
OpenAIDevs · x · 2026-09-11
Follow-up on the Agents API launch: the agent loop runs on OpenAI's infrastructure, coordinating model calls, tool use, and context, while developers control the agent's capabilities and choose where code executes and files are handled.
Related event: OpenAI Launches Agents API Public Beta with Managed Codex Harness(10 posts)→
More from coding & agent
- RD-Forget: reversible, query-dependent forgetting for agent memory — MaryamMiradi · 2026-09-11
- Vercel intern turns the terminal into an ASCII arcade for AI agents — cramforce · 2026-09-11
- Cognition Launches SWE-2 Model Matching Frontier Evals at 70% Lower Cost, Plus Devin Voice — marvinvonhagen · 2026-09-11
- My agent keeps asking me to do its job, defeating its own purpose — altryne · 2026-09-11
- stet: Open-Source Tool Lets Coding Agents Annotate the Live DOM — Funsaized · 2026-09-11
- Devs still debug agent runs for hours while demos promise autonomous research — DominiqueCAPaul · 2026-09-11