OpenAI has rolled out a set of new features for Codex, its software engineering agent, aimed at extending its usefulness well beyond a developer's own laptop. The headline addition is reusable cloud development environments that can be reached from any device. The news came during OpenAI's Dev Day on Tuesday, alongside a broader batch of Codex updates.

Among the other announcements were a refreshed Codex CLI, a new code review experience, tools for hardening infrastructure, and various API enhancements. Codex was already capable of spinning up cloud tasks to handle work remotely, but the latest move makes those cloud environments more persistent and configurable, rather than treating every cloud task as a separate, isolated remote sandbox. According to the company, developers can now run Codex wherever they need it — on a computer, remotely from a phone, or in the cloud. OpenAI says the reusable development environments are built to help tasks get started faster and to give teams a shared workspace with approved settings and permissions.

The Codex CLI, meanwhile, is gaining an update that lets developers kick off and steer tasks using their voice. A new /agents view is intended to simplify delegating work and keeping tabs on multiple tasks simultaneously. Additional changes target everyday workflows like editing prompts, resuming sessions, and working with worktrees. OpenAI is also introducing a "cleaner" terminal user interface meant to make longer sessions easier on the eyes.

Codex can also be brought into a new code review experience inside the ChatGPT desktop app. There, users can read summaries, dig into code changes, or ask Codex about possible issues before posting feedback on GitHub pull requests or GitLab merge requests. Automatic code reviews allow Codex to carry out an initial pass while users are away from their computers.

On the security front, OpenAI is introducing a set of tools called Codex Security Cloud, built to help developers protect their infrastructure. With these tools, developers can scan entire GitHub repositories on demand, on a regular schedule, or whenever new commits land. Codex then digs into the findings, removes duplicates, and prepares fixes in the cloud — even when the user is inactive and their laptop is shut. Notably, the toolset provides access to models offered through OpenAI's cybersecurity initiative Daybreak Blue, with no separate application required.

OpenAI also unveiled several API updates, among them a new Decisions API for real-time decision-making. That API relies on Luna to answer a set of user-defined questions using predefined answers. The company additionally rolled out an updated version of the Agents API that now supports computer use, along with the ability to use Amazon's Bedrock Managed Agents to build OpenAI agents that run entirely on AWS.