OpenAI introduced a new Codex desktop application for macOS, positioning it as a central control hub for developers who increasingly rely on multiple AI agents to build, ship and maintain software in parallel.
The Codex app is designed to let users supervise several agents at once, with each agent running in its own thread and organised by project. OpenAI said the tool reflects a broader shift in software development as AI models take on longer and more complex tasks that can span days or weeks, stretching beyond the limits of traditional terminals and integrated development environments.
See also: OpenAI Sets February Date to End GPT-4o Access on ChatGPT
“Models are now capable of handling complex, long-running tasks end to end,” the company said in its announcement, adding that the main challenge has shifted to how people can “direct, supervise, and collaborate with them at scale.”
Alongside the launch, OpenAI said it will temporarily include Codex with ChatGPT Free and Go plans and double usage limits for paid tiers including Plus, Pro, Business, Enterprise and Edu. The higher limits will apply across the Codex ecosystem, including the desktop app, command-line tools, IDE extensions and cloud usage.
See also: Google Rolls Out Gemini AI Features in Chrome to Boost Web Multitasking
The company also highlighted features such as built-in support for worktrees, allowing multiple agents to work on isolated copies of the same codebase, and new “skills” that extend Codex beyond code generation into research, writing and workflow automation. OpenAI said these capabilities point to a future in which developers oversee coordinated teams of agents across the full software lifecycle rather than relying on a single coding assistant.
Source: OpenAI