OpenAI Adds Reusable Cloud Workspaces to Codex for Coding Across Devices

A Codex task no longer has to begin with a fresh cloud workspace. At OpenAI’s Dev Day on September 29, the company introduced reusable development environments that developers can access from a computer or phone. The change could make remote coding work easier to resume and share, though OpenAI’s claim that tasks will start faster remains an intended benefit.

From one-off tasks to a workspace teams can reuse

Codex could already run tasks in the cloud. The new element is an environment that can persist between tasks and be configured for a team, rather than treating each remote job as a separate sandbox. Developers can reach Codex from a computer, remotely from a phone, or in the cloud.

OpenAI says the reusable environments are meant to start tasks faster and give teams a shared workspace with approved settings and permissions. Those controls are part of the design: a developer returning from another device can work in a configured environment instead of starting a new setup. Faster starts are the company’s stated goal, not a measured result in the announcement described by TechCrunch.

More ways to direct and inspect the work

OpenAI also refreshed the Codex command-line interface, or CLI—the terminal tool developers use to give Codex instructions. They can start and direct tasks by voice. A new /agents view is designed for delegating work and following several tasks at once. The update also changes prompt editing, session resumption and worktree handling, and adds a cleaner terminal display for longer sessions.

The CLI changes address the moment a developer assigns and tracks work. A separate addition addresses what happens after Codex changes code: a review experience inside the ChatGPT desktop app. Users can read summaries, examine changes and ask Codex about potential problems before giving feedback on GitHub pull requests or GitLab merge requests. Automatic reviews can also run while the user is away.

Security checks that continue after the laptop closes

The same Codex update includes Codex Security Cloud. It can scan an entire GitHub repository when requested, on a schedule, or after a new commit. Codex then investigates findings, removes duplicates and prepares fixes in the cloud. That work can continue when a developer is inactive and their laptop is closed.

Security Cloud also gives developers access to models from OpenAI’s Daybreak Blue cybersecurity initiative without a separate application. That access is a feature of this toolset, not a claim that every suggested fix has been applied. Preparing a fix still leaves a decision about whether to use it.

Taken together, the additions move Codex through more of a developer’s workflow: returning to a configured environment, directing parallel tasks, inspecting proposed code changes and preparing security fixes. The next test is whether reusable settings make that handoff genuinely easier for teams, while their review processes keep pace with work that can continue in the cloud.