Codex can now operate Windows PCs while users steer tasks from mobile
Original: Codex reaches Windows computer use with mobile steering from ChatGPT View original →
Why the Windows step matters
Codex is moving from a coding assistant that edits files toward an agent that can operate the machine where those files live. In its May 29 X post, OpenAI wrote, “Computer use now works on Windows”. That line matters because Windows remains a default development environment for many game, enterprise, .NET, embedded, and cross-platform teams.
The same post says Windows support also reaches Codex through the ChatGPT mobile app, so a user can start a task, review progress, and change direction while work continues on the PC. The tweet quickly crossed 950,000 views and more than 7,000 likes, which is strong evidence that the developer audience sees Windows support as a practical missing piece rather than a cosmetic update. OpenAI’s main account is used for first-party product and model rollouts, so this is a primary-source signal about Codex’s supported workflow.
What changes for developers
The important shift is execution context. A coding agent that can see, click, and type on Windows can test desktop apps, reproduce UI bugs, approve local flows, and inspect browser behavior without forcing the user to translate every action into text. That is especially relevant for long-running coding tasks where the agent needs to alternate between code changes and live validation.
OpenAI still frames the experience as early. The next checks are reliability, permission boundaries, latency from mobile control, and whether enterprise security teams accept remote agent access to developer machines. The source post is available on X.
Related Articles
GPT-5.6 moved from preview into access across ChatGPT, Codex and the OpenAI API. OpenAI paired the rollout with an 80.0 Coding Agent Index score, 2.8 points above Claude Fable 5, while claiming lower token use, time and cost.
OpenAI Developers posted that the Codex app is now available on Windows with a native agent sandbox and PowerShell-oriented developer environment support. The update extends Codex usage beyond previous desktop workflows and signals deeper Windows integration for agentic coding tasks.
The bottleneck in AI security is shifting from finding bugs to landing fixes. OpenAI says GPT-5.5-Cyber reached 85.6% on CyberGym, while Codex Security has scanned more than 30,000 codebases.