Imagine your AI assistant not only working for you but also keeping you company with cute animations and clear status updates. That's exactly what OpenAI's Codex pets do. These little digital buddies sit on your desktop, showing you what your AI agents are up to, and more importantly, when they need your input. The author of the original article fell in love with them and now wishes every app had such a friendly face. Let's dive into the details with this Q&A.

What exactly are Codex AI pets, and how do they help?

Codex AI pets are animated desktop companions that come with OpenAI's Codex coding app. They appear as small characters in the corner of your screen, floating above other windows. Their primary job is to show you what your AI agents are doing behind the scenes. When an agent is idling, waiting for your approval or a new task, the pet's thought bubble displays a visual cue. For example, a red clock face means the agent needs your go-ahead, while a green checkmark indicates the task is complete and it's ready for the next instruction. This keeps you informed without having to constantly switch windows or check logs. You can also drag the pet anywhere on your desktop, making it easy to keep an eye on your agents while working in other apps.

How do the visual cues work, and why are they so useful?

The visual cues are the heart of the Codex pet's usefulness. Instead of relying on notifications or manual checks, the pet's thought bubble changes to reflect the agent's status in real time. A small red clock indicates the agent is waiting for your approval before executing a command—perfect for high-stakes decisions where you want to stay in control. A green checkmark tells you the agent has finished a task and is ready for a new prompt. This is especially helpful when you're multitasking. If you're working in another window, you might miss a silent waiting agent, but the pet's persistent display—always on top of other windows—makes sure you catch it. The author notes that this feature makes them feel much more in the loop, turning a potential productivity pitfall into a smooth workflow.

Can you customize or create your own Codex pet?

Yes, and the process is straightforward. You can spawn a pet in Codex by typing the command /pet. This gives you immediate access to eight pre-made designs, ranging from the default blue 'Codex' to a duck-shaped 'Dewey,' a rock-like 'Rocky' (a nod to the movie Project Hail Mary), and even a 'BSOD' pet that humorously mimics the infamous Blue Screen of Death. To change your pet, open the Appearance menu and pick a new one. But the real fun begins when you create your own. First, install the 'hatch pet' skill via the command $skill-installer hatch-pet. Then, call the skill with $hatch-pet create a new pet inspired by my recent projects. This lets you generate a unique buddy that reflects your work or personality, adding a personal touch to your desktop.

How do Codex pets compare to similar features in other apps?

While Codex pets are charming, they aren't entirely unique. For instance, the terminal version of Claude Code includes features called 'buddies' that serve a similar purpose—keeping you informed about agent status. However, Codex pets stand out because of their visual, desktop-integrated design. They're not just text-based notifications; they live on your screen as animated characters, making them more engaging and easier to notice. The author acknowledges that there's nothing groundbreaking about the concept, but the execution makes a big difference. The pets are cute, customizable, and surprisingly useful for staying aware of background tasks. This comparison highlights a growing trend: as AI agents become more autonomous, developers are finding creative ways to keep humans in the loop without being intrusive.

Why does the author want these pets in every app?

The author's enthusiasm stems from a practical pain point. Without these pets, AI agents often wait idly for user input, especially during high-stakes tasks that require approval. If you're focused on another window, you might not notice the agent is stuck, wasting time and slowing down your workflow. Codex pets solve this by providing constant, visual status updates that stay on top of all windows. The author found that they feel much more connected to what their AI is doing, and the cute animations make the experience pleasant. Now, they wish that other agentic AI desktop apps—like Claude's Cowork application—would adopt similar companions. The bottom line: these pets are a simple but effective way to improve human-AI collaboration, making it easier to trust and delegate tasks to autonomous agents.

How do you activate a Codex pet, and what are the pet options?

Activating a Codex pet is simple. Open the Codex app and enter the command /pet in the terminal. A default pet will appear on your desktop, usually in the bottom-right corner, but you can drag it anywhere. To switch pets, go to the Appearance menu and choose from eight pre-made characters: the default blue 'Codex,' the duck-like 'Dewey,' the rock-shaped 'Rocky,' and the playful 'BSOD' among others. If you want a custom pet, you'll need to install the 'hatch pet' skill using $skill-installer hatch-pet. Once installed, you can run $hatch-pet create a new pet inspired by my recent projects to generate a unique companion. The process is designed to be creative and easy, letting you tailor the pet to your taste or current project theme.

Are Codex pets just a gimmick, or do they have real value?

While Codex pets might seem like a gimmick at first glance—after all, they're cute animations with no direct impact on code execution—their real value lies in communication and awareness. By providing continuous, visual status updates, they reduce the cognitive load of checking on your AI agents. You don't have to switch windows or parse logs; a quick glance at the pet tells you everything. The author's experience shows that this leads to a smoother workflow and a stronger sense of connection with the AI. For those who multitask heavily, this is a significant productivity boost. The author's wish for similar companions in every app reflects a broader desire for more human-centric AI interfaces. So no, they're not just a gimmick—they're a thoughtful design choice that enhances the human-AI partnership.