OpenAI’s first branded device is not the rumored smart speaker or the reported handheld, screenless hardware project. It is the Codex Micro, a $230 RGB-lit mini-keyboard built around a narrower idea: helping users keep track of multiple Codex agents from the desk.
The device turns a small set of physical keys into a status panel for agentic work. Instead of switching between windows to check whether a thread is idle, running, waiting or broken, users can look at color-coded keys and tap into the relevant Codex window when attention is needed.
What the Codex Micro is designed to do
The Codex Micro is described as a “limited-run collaboration” with Work Louder, a company that already sells a similar-looking Creator Micro line of customizable square keyboards for creative professionals. OpenAI’s version is specialized for Codex, with visible OpenAI branding on the box and on the face of the device.
Its most distinctive feature is a set of six frosted keys across the top two rows. Those keys can provide live feedback for up to six Codex threads, including threads that are not currently visible on-screen.
The color system is meant to make status readable at a glance. A key can move from white when a thread is idle, to blue when Codex is thinking, to green when a task is complete. The same keys can flash amber when Codex needs feedback or a decision from a human operator, and red when a thread hits an error.
That makes the Codex Micro less like a general keyboard and more like a small physical dashboard. A quick tap on one of the lit-up buttons brings the related Codex window onto the screen, giving the user a shortcut from status signal to action.
Why a physical Codex control panel matters
The source article frames the device around a specific working pattern: people who monitor continually running AI agents, sometimes even by keeping a half-open laptop nearby. The Codex Micro takes that monitoring behavior and compresses it into a desktop accessory.
That focus also defines its limits. A physical keyboard attachment is most useful in a desktop setting, where the user is already near the machine and can respond quickly. Away from the desk, ChatGPT’s mobile app can provide more detailed monitoring.
Still, the device points to a practical challenge created by agentic tools. When several threads are running at once, the user’s job shifts from typing every step to supervising progress, reviewing decisions and catching failures. The Codex Micro is built around that supervision layer.
The key idea is simple: make agent status visible without asking the user to keep checking the software interface. The colors do not replace the Codex window, but they reduce the friction of knowing which task needs attention first.
Controls, keycaps and function layers
Below the six frosted status keys, the Codex Micro includes six additional buttons. By default, those are mapped to common Codex tasks such as accepting changes, rejecting changes, branching threads and using a “push to talk” button for audio prompts.
Those controls are not fixed. The functions can be remapped through software, and the hardware can be changed physically with one of 32 included keycaps. That gives the device room to fit different workflows rather than only one preset layout.
The Codex layer is only the first layer. While that first customized layer is reserved for Codex, users can also program and cycle through five other function sets for general computing shortcuts.
In practice, that means the Codex Micro can serve two roles. It can be a dedicated Codex status and control surface, and it can also act as a programmable mini-keyboard when the user is doing other desktop work.
What the launch says about OpenAI hardware
The Codex Micro arrives while rumors continue around OpenAI’s broader hardware ambitions, including work on a personalized smart speaker and other devices. The source article notes that OpenAI’s interest in hardware has been apparent since at least 2023, when the company reportedly started collaborating with former Apple design chief Jony Ive.
Last May, OpenAI acquired Ive’s design firm LoveFrom. That partnership has reportedly focused on a handheld, screenless device that accepts audio and visual input from around the user.
That larger effort has also faced uncertainty. Last October, the Financial Times reported that the collaborative design work had run into unexpected technical and design hurdles that could delay the device’s reported plans for a 2026 launch. Just last week, Apple also filed a lawsuit alleging the theft by an ex-Apple employee of trade secrets related to hardware manufacturing.
Against that backdrop, the Codex Micro is a modest first step rather than a full consumer hardware debut. It is, at heart, a modified version of existing third-party hardware. But the OpenAI branding and Codex-specific software behavior make it the company’s first branded device.
Availability remains limited
The Codex Micro is taking orders now, and a spokesperson said it is expected to ship to users “shortly after purchase.” The partnership is described as a limited run, with orders accepted only “while supplies last.”
That language leaves long-term availability unclear. A representative did not respond to questions about total availability as of press time, according to the source article.
For now, the Codex Micro is best understood as a focused desktop accessory for people already working with multiple Codex threads. It does not answer every question about OpenAI’s hardware future, but it does show one concrete direction: physical tools built around supervising AI work in real time.