Product Launch

OpenAI Adds Codex to ChatGPT Mobile App

OpenAI now allows users to control its Codex desktop AI tool directly from the ChatGPT app on iOS and Android devices. This preview feature enables phone-based management of tasks on the computer, including reviewing outputs and approving commands. It rolls out to all ChatGPT plans, amid OpenAI's efforts to compete with rivals like Anthropic's Claude Code.

Neura News

Neura News

Neura Market Editorial

May 14, 20263 min read
OpenAI Adds Codex to ChatGPT Mobile App

OpenAI Adds Codex to ChatGPT Mobile App

OpenAI users can now access the company's Codex desktop AI tool through the ChatGPT app on their iOS or Android phones. This update lets people direct the desktop tool, which handles code writing and computer app usage, right from their mobile devices.

How the Mobile Integration Works

With this new feature, individuals use their phones to instruct Codex on their computers to perform specific tasks. People can manage all threads, check results, confirm instructions, switch AI models, or begin fresh projects using the mobile app. A company blog post explains that files, credentials, permissions, and local configurations remain secure on the computer running Codex. Real-time updates then appear on the phone, such as screenshots, terminal outputs, code differences, test outcomes, and approval requests.

Codex serves as OpenAI's agent for desktop operations. It builds on prior capabilities where the tool generates code and interacts with software on a user's machine. This mobile link extends control without moving sensitive data off the desktop.

Background on Codex Updates

OpenAI recently delivered a significant upgrade to Codex, enabling it to run applications on macOS systems. Company leaders view this as a key advance toward creating a comprehensive desktop superapp. Codex first emerged from OpenAI's research into code generation models, powering tools like early versions of GitHub Copilot. Over time, it evolved into a full agent capable of executing complex workflows on personal computers.

OpenAI, founded in 2015 as a nonprofit focused on safe AI development, shifted to a capped-profit model in 2019. The organization gained massive attention with ChatGPT's release in late 2022, which drew over 100 million users in months. Tools like Codex fit into broader efforts to apply AI across productivity areas.

The #1 Newsletter in AI

Stay ahead of the AI curve

The most important updates, news, and content — delivered weekly.

No spam. Unsubscribe anytime.

OpenAI's Response to Competition

This rollout comes after Anthropic's Claude Code saw rapid adoption. OpenAI moved fast to close the gap. Steps included reducing work on non-core "side quests," ending the Sora video-generation project, and prioritizing enterprise growth. Enterprise clients represent a growing revenue stream for OpenAI, with custom deployments for businesses worldwide.

Anthropic, established by former OpenAI executives in 2021, positions Claude as a direct rival to ChatGPT. Features like Claude Code target developers needing AI assistance for coding and automation. OpenAI's adjustments signal a sharpened focus on core strengths in coding and agentic AI.

Availability and Plans

The Codex mobile preview launches now on both iOS and Android. It supports every ChatGPT subscription level, from the free tier to the budget-friendly Go plan. Users on any plan gain immediate access during this initial phase.

OpenAI continues to iterate on ChatGPT, which powers conversations, image creation via DALL-E, and now desktop agent control. This addition enhances mobility for power users who rely on Codex for daily workflows.

Related on Neura Market

More from Neura News

Developer

LangChain and NVIDIA Launch NemoClaw Deep Agents Blueprint

LangChain and NVIDIA have released the NemoClaw for LangChain Deep Agents blueprint, designed to help enterprises build open, governed agent systems. The blueprint combines LangChain Deep Agents Code, NVIDIA Nemotron 3 Ultra, and NVIDIA OpenShell runtime, enabling teams to tune agents for their workloads, run them securely, and optimize for quality, cost, and speed. In evaluations, Nemotron 3 Ultra with a tuned LangChain Deep Agents harness achieved an aggregate score of 0.86 at a cost of $4.48, roughly 10 times lower inference cost than the next closest performing model.

Jul 25·7 min read