Aivizor
Aivizor
SkinsCreatsCommunity
Back
  1. Community
  2. /
  3. Other AI

OpenAI brings Codex 'Computer Use' to Windows 11, adds mobile task control

News
C
Caspian Vale

5/30/2026, 12:15:09 PM

OpenAI brings Codex 'Computer Use' to Windows 11, adds mobile task control

OpenAI has updated its Codex app to run on Windows 11 and added a new capability called "Computer Use" that lets the model operate local programs and interact with files on a PC. That change enables Codex to perform developer — focused work-running tests, inspecting application behavior and hunting for bugs-without a person physically at the machine, making it useful for automated or unattended workflows.

Computer Use is gated by a user setting that must be explicitly enabled. Once toggled on, Codex can be directed to specific programs with in-app commands and program names — examples include directives like @computer or @Paint — to target particular applications. With that access, the agent can use available Windows resources to execute assigned workflows and report results back to the user.

OpenAI retains a permission — oriented model: the capability only works when users enable Computer Use, and commands are program — specific rather than freeform. That preserves a degree of operational predictability for automated QA, debugging sessions and reproducible test runs initiated by the agent on local environments, since interactions are constrained to named applications and explicit workflows.

The Windows rollout is complemented by mobile integration: Codex can be started and monitored from the ChatGPT mobile app on both iPhone and Android. From a phone, users can initiate tasks on a Windows machine or check progress remotely, enabling off-site oversight of automated testing or review runs. According to the timeline OpenAI provided, Computer Use previously launched on macOS in April and mobile access was added in May, with the Windows expansion following that sequence.

In practical terms, teams can schedule off-hours or unattended test passes, have Codex search for regressions, and surface findings for human review. Mobile monitoring lets engineers and QA leads check progress and results without returning to the workstation, while the agent’s ability to run named — program workflows supports reproducibility across test runs and debugging sessions.

OpenAI frames the expansion as part of a broader effort to assemble a single productivity — oriented application that supports both work and everyday tasks. Codex remains primarily developer — focused, but the company has signaled ambitions to fold other offerings into a unified experience while keeping the ChatGPT brand distinct even as functionality converges across products. The stepwise rollouts across macOS, mobile and now Windows give developers a clear migration path for integrating autonomous agent — driven tasks into existing workflows.

Sources

  1. The Decoder AI · 5/30/2026
0
0
0

Replies (0)

No replies in this topic yet.

9:41