OpenAI has launched a new desktop application for its Codex AI coding system, aiming to enhance software development on macOS. The app allows developers to manage multiple AI coding agents simultaneously, transforming the coding process from a single assistant model to a multi-agent system. This innovation enables developers to automate repetitive tasks and manage long-running processes, with agents operating independently for up to 30 minutes.
OpenAI’s Codex app is introduced at a crucial time for the enterprise AI market. A recent survey by Andreessen Horowitz highlights that 78% of enterprise CIOs are using OpenAI models in production, but competitors like Anthropic and Google are gaining traction. Anthropic has notably increased its enterprise presence by 25% since May 2025. OpenAI’s new app is strategically positioned to maintain its lead in the highly competitive AI coding tools segment.
The release of the Codex app signals a shift in how developers engage with AI in coding. Unlike previous tools that focused on real-time code completion, this application allows for managing multiple agents and executing parallel tasks. OpenAI’s CEO, Sam Altman, emphasized the app’s ability to handle complex coding projects without traditional development environments. The app’s capabilities include skills for connecting to tools and running workflows, along with automations that allow for background task scheduling.
As OpenAI continues to innovate, the company plans to expand the Codex app to Windows and enhance its features, including support for cloud-based triggers and customizable personalities. This launch is part of OpenAI’s broader strategy to establish Codex as a leading tool for AI-assisted development, amid a rapidly growing market for enterprise AI solutions.
