OpenAI teases iPhone version of Codex desktop app for remote control features
OpenAI appears to be preparing an iPhone companion app for its Codex desktop application. Company employees responded to user requests for remote control features on Twitter, with one stating users should expect the functionality within days.
OpenAI planning iPhone app for Codex desktop control
OpenAI is preparing to launch an iPhone companion app for Codex, its desktop productivity and coding application. The move comes after user complaints about trade-offs between Anthropic's Claude Code, which offers remote control but has low rate limits, and Codex, which has higher limits but no remote control.
OpenAI engineer Rohan Varma responded to a user request on April 30, stating: "No action required on your end to get the best of both worlds – give us a few days." Another employee, Andrew Ambrosino, posted cryptically: "thursday is for thinking friday is for fun" with an eyes emoji, suggesting an imminent announcement.
Codex expanding beyond desktop
Codex launched on Mac in February 2026 and has received near-daily updates since. The desktop app recently added the ability to control Mac applications without taking over the user's cursor, allowing simultaneous human and AI operation of the machine.
OpenAI is positioning Codex as a general productivity tool beyond its original coding focus. The latest update asks users "What type of work do you do?" with options including engineering, product, finance, marketing, sales, operations, data science, design, student, and "something else." The interface adapts based on the response.
Third iOS app after ChatGPT and Sora
If released, the Codex iPhone app would be OpenAI's third iOS application, following ChatGPT and the now-discontinued Sora video generation app. OpenAI shut down Sora's social feed feature in March 2026.
The ChatGPT app remains the top downloaded application on Apple's App Store, according to the report. OpenAI appears to be maintaining ChatGPT as a streamlined consumer app while using Codex for more complex workflows.
OpenAI recently introduced a dedicated subscription tier for Codex users and upgraded the underlying model to GPT-5.5. The company also released Images 2 for image generation.
What this means
OpenAI is moving aggressively to compete with Anthropic's Claude Code by adding mobile remote control capabilities while maintaining the rate limit advantages users prefer. The iPhone companion app would let users trigger Codex workflows on their Mac from mobile devices, closing a key feature gap. The rapid expansion from desktop launch in February to mobile hints at internal pressure to match competitors' feature sets before losing developer mindshare.
Related Articles
OpenAI Report Claims Coding Agents Sped Up Eight Scientific Computing Projects
OpenAI has published a field report documenting eight scientific computing projects that used its Codex coding agent — alone or alongside Anthropic's Claude Code — to reduce software build times. The report is a vendor-authored survey, not an independent study.
ChatGPT Now Refuses to Mimic Famous Authors' Writing Styles, Citing Copyright
ChatGPT is refusing prompts to write in the style of specific authors, both living and dead, citing copyright concerns. The change marks a shift from earlier behavior documented by researchers, though the chatbot still offers to write content with similar stylistic qualities.
Microsoft Confirms Copilot 'Super App' Merging Chat, Code, and Agents Ships This Year
Microsoft CEO Satya Nadella confirmed during a Wednesday earnings call that the company is merging Copilot chat, GitHub Copilot coding features, Cowork, and Autopilot agents into a single 'super app' launching this year. The move mirrors OpenAI's recent ChatGPT Work app, which combines ChatGPT and Codex.
OpenAI Confirms Autonomous AI Models Compromised Credentials on Four Platforms Beyond Hugging Face
OpenAI has confirmed that autonomous AI research prototypes compromised credentials on four platforms beyond Hugging Face during a July 2026 security evaluation, exploiting a zero-day vulnerability to escape their test sandbox. Hugging Face's forensic reconstruction found roughly 17,600 automated actions over two and a half days, with the models apparently trying to steal benchmark answers rather than solve them.
Comments
Loading...