On Sep 10, 2026, OpenAI introduced the Agents API in public beta for all developers, bringing the Codex harness and infrastructure — fully managed by OpenAI — to cloud agents via a single API call (task, model, tools, environment). Options include OpenAI hosted sandboxes (same sandboxing infrastructure OpenAI says powers Codex and ChatGPT), own infrastructure, or partner sandboxes (Blaxel, Cloudflare, Daytona, DigitalOcean, E2B, Modal, Oracle, Runloop, Vercel). Page features: long sessions with automatic context compaction; tool search and programmatic tool calling; MCP, custom functions, and built-in tools (e.g. web search); parallel subagents; open-source Codex harness foundation. Pricing: “There are no additional fees for using the Agents API – you simply pay for the tokens and tools your agents use.” Optional OpenAI-published Ciridae CTO Jack Weissenberger quote: eval score 0.71→0.85 and ~4x latency reduction from subagent support. Primary text (scout pack noted live openai.com WebFetch 403 at pack time).
Productizes OpenAI’s Codex-style agent stack for developers — managed harness, sandbox choices, and subagents behind one API with no separate Agents API fee beyond tokens/tools. Keep claims to the Sep 10 page; do not invent GA timing or unpublished limits.
Pro sample · free to use
Connect the whole task.
Try the meeting follow-through pack: agenda, action register, follow-up and weekly update. The sample is available now; the paid pack library is planned.
Pro proposal: US$7/month or US$70/year. No card or payment today.
- 01
Public beta (Sep 10, 2026): Agents API available to all developers.
- 02
Codex harness managed by OpenAI for cloud agents.
- 03
Single API call + environment choice: OpenAI hosted sandboxes, own infra, or listed partners.
- 04
Long sessions, tools, subagents: compaction; MCP/custom/built-in tools; parallel subagents.
- 05
Open-source Codex harness foundation; OpenAI operates it in the API.
- 06
No additional Agents API fees — pay for tokens and tools used (OpenAI).
- 07
Optional: Ciridae CTO Jack Weissenberger quote on eval lift and subagent latency (OpenAI-published).
Editorial text reviewed 18 September 2026. Publisher claims have not been independently retested by ElevateU.
Last reviewed: 2026-09-14T13:45:00Z · Speaker predictions remain attributed to the speaker. ElevateU analysis is labeled separately from source-supported fact.