OpenAI on February 2 launched the Codex app for macOS, introducing a desktop interface to help developers manage multiple AI agents, run tasks in parallel, and supervise long-running software projects.
The Codex app is available starting February 3 on macOS. OpenAI said usage is included with eligible ChatGPT subscriptions, with additional credits available for purchase.
The company said the new app acts as a “command centre for agents”, allowing developers to delegate work across multiple threads, review and comment on code changes, and switch between tasks without losing context. The app supports isolated worktrees, enabling several agents to work on the same repository without conflicts.
“Models are now capable of handling complex, long-running tasks end to end, and developers are now orchestrating multiple agents across projects,” OpenAI said, adding that the challenge has shifted to “how people can direct, supervise, and collaborate with them at scale.”
Alongside the launch, OpenAI announced expanded access to Codex. For a limited time, Codex will be included for ChatGPT Free and Go users, while rate limits for Plus, Pro, Business, Enterprise, and Edu plans have been doubled. The higher limits apply across the Codex app, CLI, IDE extensions, and cloud usage.
The company said the Codex app is aimed at supporting a new way of building software, ranging from pairing with a single agent on small edits to overseeing coordinated teams of agents across design, development, testing, and maintenance. “Existing IDEs and terminal-based tools are not built to support this way of working,” it added.
The app has introduced Skills, which bundle instructions, resources, and scripts so Codex can reliably use external tools and workflows. OpenAI said skills allow Codex to go beyond code generation into tasks such as information gathering, document creation, image generation, deployment, and project management.
Developers can explicitly invoke skills or allow Codex to select them automatically based on the task.
OpenAI said it has used Codex internally for tasks such as running evaluations, drafting documentation, managing issue backlogs, and generating reports. The app includes a skills library and allows teams to share custom skills across projects.
Another feature, Automations, lets users schedule Codex to run tasks in the background on a defined cadence. Results are delivered to a review queue once a run is complete. OpenAI said it uses Automations for activities such as daily issue triage and monitoring CI failures.
The company said developers can also choose between two interaction styles—a terse, execution-focused personality or a more conversational one—using a command available across the app, CLI, and IDE extensions.
On security, OpenAI said the Codex app uses system-level sandboxing similar to the Codex CLI. By default, agents are restricted to specific folders or branches and must request permission for elevated actions such as network access, with configurable rules available for teams.
Looking ahead, the company said it plans to release a Windows version, expand automation capabilities with cloud-based triggers, and continue improving multi-agent workflows.
ALSO READ: GitHub Introduces Copilot SDK to Embed AI Agents in Applications
