Introducing the Codex app: A Command Center for Agents

Since the launch of Codex in April 2025, the way developers work with agents has fundamentally changed. Models are now capable of handling complex, long-running tasks end to end and developers are now orchestrating multiple agents across projects: delegating work, running tasks in parallel, and trusting agents to take on substantial projects that can span hours, days, or weeks.

The new multi-agent sidebar allows for seamless context switching between different project threads.

The core challenge has shifted from what agents can do to how people can direct, supervise, and collaborate with them at scale. Existing IDEs and terminal-based tools are not built to support this way of working. This new way of building coupled with new model capabilities demands a different kind of tool, which is why we are introducing the Codex desktop app, a command center for agents.

The Codex app provides a focused space for multi-tasking with agents. Agents run in separate threads organized by projects, so you can seamlessly switch between tasks without losing context. The app lets you review the agent’s changes in the thread, comment on the diff, and even open it in your editor to make manual changes.

Individual threads allow you to supervise specific tasks while other agents work in the background.

This parallel approach means you no longer have to wait for a single agent to finish a large refactor before starting work on a new feature. You can have one agent triaging bugs in your stable branch while another explores a new architectural pattern in a feature branch. The interface handles the heavy lifting of context management, ensuring that each interaction is tailored to the specific agent's current task and environment.

One of the most powerful features of the Codex app is its built-in support for worktrees. This allows multiple agents to work on the same repository without conflicts. Each agent works on an isolated copy of your code, allowing you to explore different paths without needing to track how they impact your local codebase.

As an agent works, you can check out changes locally or let it continue making progress without touching your local git state. This solved a major pain point for developers: the fear of an agent "corrupting" the current working directory while a manual task was in progress. With worktrees, the agent's environment is truly separate yet perfectly synchronized with the repository's history.

Codex is evolving from an agent that writes code into one that uses code to get work done on your computer. With skills, you can easily extend Codex beyond code generation to tasks that require gathering and synthesizing information, problem-solving, writing, and more.

The Skill Library centralizes reusable agentic capabilities across your entire team.

Skills bundle instructions, resources, and scripts so Codex can reliably connect to tools, run workflows, and complete tasks according to your team’s preferences. The Codex app includes a dedicated interface to create and manage skills. You can explicitly ask Codex to use specific skills, or let it automatically use them based on the task at hand.

Case Study: Voxel Velocity

To demonstrate the power of skills, we asked Codex to make a racing game, complete with different racers, eight maps, and even items players could use with the space bar. Using an image generation skill and a web game development skill, Codex built Voxel Velocity by working independently using more than 7 million tokens with just one initial user prompt.

It took on the roles of designer, game developer, and QA tester to validate its work by actually playing the game. This level of autonomy is only possible because the agent had access to high-level skills that abstracted away the boilerplate of game engine setup and asset pipeline management.

Core Skills Included at Launch:

With the Codex app, you can also set up Automations that let Codex work in the background on an automatic schedule. Automations combine instructions with optional skills, running on a schedule you define. When an Automation finishes, the results land in a review queue so you can jump back in and continue working if needed.

Schedule recurring tasks like issue triage, release summaries, and test runs.

At OpenAI, we have been using Automations to handle repetitive but important tasks: like daily issue triage, finding and summarizing CI failures, generating daily release briefs, checking for bugs, and more. This shifts the developer's role from executing these tasks to reviewing the synthesized output, significantly increasing team velocity.

Developers have different preferences in how they work with an agent. Some want a blunt, execution-focused partner; others prefer more communicative, engaging interactions. Codex now lets developers choose between two personalities, a terse, pragmatic style and a more conversational, empathetic one, without any change in capabilities, to fit the approach you like the most.

Choose between Terse and Conversational styles to match your preferred workflow.

Secure by Default, Configurable by Design

We are integrating security by design across the entire Codex agent stack. The Codex app uses native, open-source and configurable system-level sandboxing just like in the Codex CLI. By default, Codex agents are limited to editing files in the folder or branch where they are working and using cached web search, then asking for permission to run commands that require elevated permissions like network access.