OpenAI adds Codex remote control to ChatGPT mobile

9to5mac reports that OpenAI is adding remote access to Codex inside the ChatGPT mobile app for both iPhone and Android. 9to5mac quotes OpenAI saying "Codex is now in the ChatGPT mobile app so you can stay in the loop from anywhere while Codex gets work done across your laptops, devboxes, or remote environments." 9to5mac reports the mobile integration lets users work across active threads to review outputs, approve commands, change models, or start new tasks while Codex runs on a connected machine. 9to5mac reports the company describes updates flowing back to the phone in real time, including screenshots, terminal output, diffs, test results, and approvals. 9to5mac reports the rollout begins today for iPhone and Android and that the article says Windows support will follow.

9to5mac reports OpenAI describes the mobile app loading the live state from any environment where Codex is running so users can "work fluidly across active threads, approvals, plugins, and project context." 9to5mac reports the article states files, credentials, permissions, and local setup remain on the machine where Codex operates, with the mobile app receiving outputs and artifacts in real time. 9to5mac reports OpenAI frames the feature as more than single-task remote control, enabling cross-thread supervision and new prompts from a phone while work continues on the host machine.

Editorial analysis: Technical context: Industry observers note that adding mobile supervision to long-running agentic systems typically raises demands for secure remote authentication, session isolation, and tamper-evident audit trails. Editorial analysis: Context and significance: For practitioners, on-device control and real-time artifact streaming change operational tradeoffs between local compute, remote management, and observability; teams that adopt similar patterns often need to formalize approval workflows and integrate mobile-friendly alerting. Editorial analysis: What to watch: For practitioners, monitor how OpenAI and downstream users handle authentication for remote access, what audit data is retained, and whether plugins or third-party integrations are supported in mobile-supervised sessions.