Codex launches into the ChatGPT mobile app, enabling remote control of Mac development tasks

ChainNewsAbmedia

OpenAI announced on May 14 that it is integrating Codex into the ChatGPT mobile app. 9to5Mac reports that the new feature is rolling out on iOS, iPadOS, and Android in a preview version, and all ChatGPT plans (including Free and Go) can use it. The mobile version does not execute code; instead, it acts as a remote control interface, connecting to the macOS Codex desktop version or a remote development environment via SSH.

Phone approval of tasks, viewing diffs, switching models, and real-time terminal output

The Codex mobile app focuses on directing and reviewing development tasks. Users can start new tasks from their phone, view all ongoing conversations, approve agent execution commands, switch models, and add context; when Codex runs tasks on the macOS side, the phone can receive terminal output in real time, along with screenshots, code diffs, and test results. OpenAI emphasizes that files, credentials, permissions, and local environment settings remain on the executing device, with only status updates and results synced to the phone.

At present, the mobile app can only connect to the macOS version of the Codex desktop app (Mac and Mac mini). The Windows version “is coming soon,” but no specific timeline has been announced. The Codex desktop app is a local development agent released by OpenAI in February 2026. This outlet previously reported that on May 13, Codex added Symphony multi-agent and Computer Use cross-application functionality.

Introduced in parallel: Remote SSH and Hooks official versions—Codex connects remote dev environments and custom scripts

OpenAI is also upgrading two existing preview features to official versions. Remote SSH enables Codex to connect to enterprise-approved remote development environments, reusing that environment’s dependency packages and security policies, and the desktop app can also directly read the user’s existing SSH configuration files to automatically detect available hosts. Hooks lets developers insert custom scripts into Codex’s workflow loop; common use cases include scanning keys in prompts, running validators, recording conversations, creating memory entries, and tailoring behavior for specific software projects or directories.

Programmatic access tokens are only available for Enterprise and Business plans. Codex deployments that comply with HIPAA are limited to qualified ChatGPT Enterprise workspaces that have local environment access.

Codex interface expansion: mobile approval, desktop execution, and Chrome testing run in parallel

This mobile rollout is a key extension following the release of the desktop version at the end of February. Over the past two months, OpenAI has continued to expand Codex’s access interfaces. It first rolled out a Chrome extension that allows web app testing inside the browser and collects cross-page context. The desktop app added a desktop pet and the multi-agent Symphony mode, while the mobile app completes the last piece of “you can keep reviewing AI development agents even when you’re away from your computer.” All three interface versions share the same Codex backend and local file/permissions system, allowing users to switch seamlessly between devices.

This article about Codex coming to the ChatGPT mobile app for remote control of Mac development tasks first appeared on Lian News ABMedia.

Disclaimer: The information on this page may come from third parties and does not represent the views or opinions of Gate. The content displayed on this page is for reference only and does not constitute any financial, investment, or legal advice. Gate does not guarantee the accuracy or completeness of the information and shall not be liable for any losses arising from the use of this information. Virtual asset investments carry high risks and are subject to significant price volatility. You may lose all of your invested principal. Please fully understand the relevant risks and make prudent decisions based on your own financial situation and risk tolerance. For details, please refer to Disclaimer.
Comment
0/400
No comments