OpenAI Adds Codex Remote Access to ChatGPT Mobile Apps

OpenAI has announced remote access to its Codex system through an update to the ChatGPT app for iOS, extending the capability to iPhone users. The feature brings control of Codex for Mac directly into the ChatGPT mobile application rather than requiring a separate app.

Codex on mobile lives inside the ChatGPT app. While Codex operates as a standalone app on the Mac, OpenAI has integrated remote access features into the ChatGPT app for iPhone. This allows users to stay in the loop from anywhere while Codex continues working across laptops, devboxes or remote environments.

The company points to practical uses including approving Codex tasks and starting new prompts from a phone when away from a computer. As agents take on longer-running work, OpenAI notes that a new rhythm for collaboration is emerging. Users need to easily answer questions, review what Codex found, change direction, approve what comes next or add a new idea.

According to OpenAI, Codex in the ChatGPT mobile app delivers a fully-featured mobile experience. When connecting to any machine where Codex is running — whether a laptop, dedicated Mac mini or managed remote environment — the app loads the live state from that environment. This enables fluid work across active threads, approvals, plugins and project context.

OpenAI describes the tools as more than basic remote control. From a phone, users can work across all threads, review outputs, approve commands, change models or start something new. Files, credentials, permissions and local setup remain on the machine where Codex operates. Updates flow back to the phone in real time, including screenshots, terminal output, diffs, test results and approvals.

The Codex remote access through the ChatGPT mobile app is rolling out starting today for both iPhone and Android. OpenAI says support for remotely controlling Codex for Windows will follow. The company shared that the feature is now in preview, allowing users to start new work, review outputs, steer execution and approve next steps while Codex keeps running on a laptop, Mac mini or devbox.