OpenAI Codex Comes to the ChatGPT Mobile App

OpenAI has added its Codex coding agent to the ChatGPT mobile app in a new preview for iOS and Android, giving developers a way to monitor and steer coding tasks running on other machines directly from their phone. Instead of turning the phone into a full development environment, the mobile experience works as a remote control for Codex sessions running on laptops, Macs, or remote development boxes.

This approach lets long‑running coding jobs continue in the background while developers check progress, review changes, or approve actions from anywhere.

What Codex Mobile Actually Does

Codex is OpenAI’s AI agent designed to perform software‑development tasks such as writing code, refactoring projects, running commands, and executing tests across repositories.

The new mobile preview extends that workflow by connecting the ChatGPT app to active Codex environments. When a developer signs into the same ChatGPT account on both devices, the phone can display and control sessions running on the connected host machine.

Rather than transferring files to the phone, the mobile app simply interacts with the existing environment where the code and credentials already live.

Remote Control for Coding Sessions

The core idea behind the mobile release is that Codex keeps running on the computer, while the phone becomes a control surface.

Developers can:

• View active Codex sessions and environments
• Check whether tasks are running, waiting, or blocked
• Send new instructions or adjust the agent’s plan
• Approve commands that require confirmation
• Start or redirect tasks remotely

Because the coding agent continues running on the original machine, the project’s files, tools, and credentials remain in their existing development environment.

Key Features in the Mobile Experience

The mobile interface focuses on monitoring and steering work rather than writing code directly. Early previews highlight several core capabilities.

Live session status

The ChatGPT mobile app shows the current state of Codex tasks, including whether the agent is actively working, waiting for input, or blocked by an approval step. This helps developers track progress without keeping a laptop open.

Command approvals and unblocking

Many coding agents pause when they need confirmation for actions such as running commands or modifying important files. The mobile interface allows developers to approve or reject these actions from their phone, preventing workflows from stalling.

Terminal and output monitoring

Users can follow the agent’s activity through terminal output, logs, screenshots, test results, and code diffs generated by the running environment. This visibility lets developers review what the agent is doing in real time.

Steering and model control

Developers can send additional instructions or adjust how Codex proceeds with a task. Some reports also describe options to switch models or modify execution settings from the mobile interface, although detailed UI behavior may vary in the preview.

Project context and plugins

Codex operates within a selected project directory on the connected machine, preserving the repository structure, tools, and context used during development. The broader Codex ecosystem also includes plugins and integrations with developer tools, which remain available within the host environment.

Why Developers Asked for Mobile Access

AI coding agents often run tasks that take minutes or even hours—especially when they are analyzing large repositories, generating new features, or running tests.

Before mobile access, developers frequently had to leave a laptop nearby to monitor progress or approve steps. The new mobile integration addresses that problem by letting them check and manage their agent from anywhere, keeping workflows moving even after leaving their desk.

This capability also fits the broader trend toward agent‑driven development, where AI systems perform larger chunks of work autonomously while humans supervise and intervene when necessary.

Who Can Use Codex on Mobile

The Codex mobile feature is currently rolling out as a preview inside the ChatGPT app on iOS and Android. Reports indicate that the capability is available across ChatGPT subscription tiers, including Free and Go plans, although advanced model access or higher usage limits may still depend on higher‑tier subscriptions.

The Bigger Picture

Bringing Codex to mobile connects another piece of OpenAI’s broader developer workflow. Codex already operates in multiple places—including terminals, IDE extensions, and dedicated apps—linked through a single ChatGPT account.

Adding a mobile control layer means developers can start tasks on their main machine and then monitor, steer, and unblock them from anywhere, turning AI coding agents into something closer to always‑running collaborators rather than tools that only work when a laptop is open.