OpenAI adds Codex to ChatGPT mobile app for on-the-go coding

At a glance:

• OpenAI integrates Codex into ChatGPT mobile app for remote coding control
• Available now on iOS and Android across all plans including Free tier
• Users can monitor, approve, and direct coding projects from phones while files remain on local machines

What happened

OpenAI announced this week that it's bringing its programming tool Codex directly into the ChatGPT mobile application. This significant update allows developers to stay connected to their coding projects even when they're away from their primary machines. The feature enables users to monitor progress, review outputs, approve commands, and direct their AI coding assistants from their smartphones, eliminating the need to carry laptops everywhere just to maintain continuous coding workflows.

The integration represents a shift in how developers interact with AI coding tools. Previously, users needed to keep laptops open and connected to Wi-Fi to maintain their AI agents' operations. With this new mobile integration, developers can now respond to coding needs, review what Codex has found, change directions, approve next steps, or add new ideas directly from their phones, regardless of their physical location. This flexibility could transform how developers work in various environments, from coffee shops to airports to their children's activities.

How it works

The Codex integration works by connecting the mobile app to the machine where Codex is actually running, whether that's a laptop, desktop, development box, or remote location. The mobile app loads the live state from the machine, allowing users to review outputs, approve commands, change models, and monitor current tasks or create new ones across all threads. Importantly, customers' files, credentials, permissions, and local setup remain securely on the machine where Codex is running, not on the mobile device.

OpenAI has ensured security through a secure relay layer, meaning machines running Codex remain protected and not exposed to the public internet. Users can view screenshots, terminal outputs, and other information directly on their phones while maintaining the security of their development environment. The company emphasizes that this architecture protects sensitive code and credentials while providing the flexibility of mobile access.

Why it matters

This mobile integration addresses real pain points for developers who have become dependent on AI coding assistants. The article mentions several people who had to go about their daily lives with laptops open, even if just a crack, to keep Codex running and writing code. OpenAI even acknowledged this phenomenon with a humorous video on TikTok. The new feature liberates developers from this tether, allowing them to accomplish tasks without juggling a laptop or having to perch it somewhere inconvenient.

In practical terms, the mobile app enables developers to respond to coding needs as they arise in real-world situations. For example, if Codex identifies a bug while a developer is at the grocery store, the agent can detect the issue, test for it, and start fixing it. Or if Codex needs a decision while a developer is socializing at a coffee shop, they can check the situation and direct the appropriate action. Even spontaneous ideas that pop up during other activities can be immediately sent to Codex for implementation, creating a more fluid and responsive development workflow.

The future

OpenAI is rolling out the feature in preview on iOS and Android across all plans, including Free and Go, in all supported regions. The company has also announced that support for Windows is coming soon, though specific timelines weren't provided. To try out the new functionality, users will need to update both the ChatGPT mobile app and the Codex app on macOS.

With more than 4 million people using Codex every week, this mobile integration has the potential to significantly impact how developers interact with AI coding tools. As AI agents become more capable of handling complex coding tasks independently, the ability to monitor and direct these processes from anywhere could become increasingly valuable. The integration of AI coding tools into mobile workflows represents another step toward making AI assistance ubiquitous and accessible across all aspects of professional life.