CodexBar is a compact macOS 14+ menu bar app that displays usage quotas, sessions, and weekly reset times for various AI tools such as Codex, Claude, Cursor, Gemini, Copilot, and more in real time, supports dynamic icons, shortcut switches, and uses local processing to protect privacy.
You can download it from GitHub or Homebrew, enable the corresponding service platform in the settings, and complete the authorization through cookies or CLI to use it.
It helps you avoid accidentally triggering token limits, track usage quotas without interrupting workflows, and efficiently control usage costs.
The way developers work has changed a lot over the past year. Increasingly, code is no longer entirely handwritten by humans, but is being done collaboratively through various AI tools. From writing functions and explaining code to automatically generating scripts, AI has become a part of many developers’ daily routines.
Problems also arise.
Many people use multiple AI tools at the same time, such as Claude, GitHub Copilot, Cursor, Gemini , etc. Each service has its own quota limit, some are calculated by tokens, some are limited by the number of requests, and some are reset on a weekly basis. When you’re immersed in your workflow, it’s easy to overlook these limitations and not realize the problem until you suddenly encounter the “Quota exhausted” prompt.
This is exactly what CodexBar wants to solve.
CodexBar is a very small macOS menu bar app that doesn’t try to replace any AI tools or change your development process. What it does is simple: it centralizes the usage of various AI services in the system menu bar, allowing developers to keep track of their usage limits.
Once installed, CodexBar will remain quiet in the macOS menu bar. With simple authorization settings (such as cookies or CLI login methods), it can read the usage status of different AI platforms and display the current quota status in the menu bar. If you’re using a service like Claude, Copilot, or Cursor, it will tell you how much quota you have left and when the next quota will reset.
This design is actually very much in line with macOS habits. The menu bar is originally a place to place system status information, such as power, network, CPU usage, etc. CodexBar simply turns “AI usage quotas” into a new system status information. Developers can see current AI usage at a glance without having to open a web page or console.
Another interesting detail is the dynamic change of the icons. CodexBar adjusts the menu bar icons based on the status of different services, such as alerting you when you’re close to the usage limit. In this way, even if you don’t open the menu, you can quickly determine whether the current quota is sufficient.
In terms of privacy, this tool also adopts a more restrained design. All data processing is done locally, it does not proxy your AI requests, and it does not upload usage data to third-party servers. It acts more like a native monitoring tool than an intermediate layer of an AI platform.
In a sense, CodexBar reflects a new trend. In the past, developers needed to monitor CPU, memory, and disk space, but now, AI usage quotas have become a resource that needs to be managed. When AI becomes a daily development tool, tokens and calls need to be monitored and managed, just like server resources.
CodexBar has no complex features and no grand goals. It simply organizes a piece of information that was originally scattered across different platforms into a simple and clear menu bar state. This minimalist design instead makes it very practical.
When AI becomes part of the development process, perhaps in the future the operating system menu bar will show not only battery and network, but also your AI quota. What CodexBar does is bring this future to the present.
Github：https://github.com/steipete/CodexBar
Tubing: