CodexBar: a macOS menu-bar view of Codex and Claude Code usage
Show usage stats for OpenAI Codex and Claude Code, without having to login.
At a glance
- What is it?
- CodexBar is an MIT-licensed macOS menu-bar app that shows usage and spending stats for OpenAI Codex, Claude Code and other AI coding tools without a separate login, reading local data. It puts your rate-limit windows and costs a glance away.
- Who is it for?
- Adopt CodexBar if you use AI coding tools like OpenAI Codex and Claude Code on macOS and want their usage windows and spending visible in the menu bar without a separate login. Do not expect it on other platforms or to manage your accounts, and treat its spending figures as estimates from local data rather than authoritative billing.
- Can I use it commercially?
- Yes. MIT is a permissive licence: you can use, modify and sell software built on it, as long as you keep its copyright and licence notices.
- Is it still maintained?
- Yes. The repository last received commits 1 day ago.
- What is it written in?
- Mainly Swift, according to GitHub's language statistics.
Answers come from the project's GitHub data, last synced on September 29, 2026, and from our analysis. They are not legal advice.
DEEP OPEN-SOURCE ANALYSIS
What CodexBar shows
Developers who lean on AI coding tools keep hitting the same question: how much of my usage window or budget is left? CodexBar answers it from the menu bar. It is a macOS app that shows usage stats for OpenAI Codex and Claude Code, and other AI tools, without requiring you to log in separately, reading the data those tools already have locally. It surfaces rate-limit windows, spending totals and reset timing so the information is a glance away rather than something you fetch from a dashboard. The audience is heavy users of Codex, Claude Code and similar tools on macOS who want ambient awareness of their limits and costs. It is a native Swift menu-bar utility, so its role is presentation: it consolidates usage and spend across your AI coding tools into one always-visible place, using data already on your machine rather than new credentials.
Reading local usage without a new login
The mechanism is reading the usage and account data your tools store locally and presenting it, without asking you to authenticate again. Because Codex, Claude Code and the others keep track of their own usage, CodexBar surfaces that rather than logging in as you, which is the no-login promise, and it combines multiple tools' figures into one menu-bar view with spending totals and reset windows. Its changelog shows attention to correctness across providers, keeping different tools' allowances tied to their own reset windows and assembling more complete spending totals, which matters because each provider counts and resets differently. It is a native macOS app with a widget extension and an appcast for updates, so it fits the platform. The design is a lightweight, always-present readout: it does not change how you use the tools, it just makes their usage and cost legible in one place, continuously.
Installing CodexBar
CodexBar is distributed for macOS through Homebrew, which is the simplest install:
brew install --cask codexbarThere is also a tap-based formula, and the README lists other package routes. Once installed, it lives in the menu bar and reads the local data from your AI coding tools to display usage, so there is no separate sign-in step to configure. The first real use is installing it, opening the menu-bar item, and confirming it shows your current usage windows and spending for the tools you use, Codex, Claude Code and others it supports, which verifies it is reading your local data correctly. Because it distributes through Homebrew with an appcast, keeping it current is straightforward. It is a small utility, so setup is minimal: install, glance, and it is doing its job.
Where a usage readout is bounded
The limitations are those of a convenience utility. CodexBar is macOS-only, so it does nothing for other platforms, and it is a viewer over data the tools already keep, so it depends on those tools being installed and on the format of the data they store, which means a change on a tool's side can affect what it can read. Spending figures are derived from that local data, so treat them as close estimates of usage and cost rather than authoritative billing. It shows and consolidates information; it does not manage your accounts or change your limits. And because each provider structures usage and resets differently, accuracy across the growing set of supported tools depends on the app keeping pace, which its frequent releases suggest it does. These are the ordinary boundaries of a menu-bar readout, and they mean CodexBar is for awareness, not for account control or exact accounting.
CodexBar versus provider dashboards or CodexIsland
The alternatives are checking each provider's own dashboard, or a similar surface such as a notch-based usage viewer. Provider dashboards are authoritative but require you to open each one, and they do not consolidate multiple tools or sit in front of you. A notch-based viewer occupies the MacBook notch instead of the menu bar and covers a similar goal of ambient AI usage. CodexBar's difference is that it is a menu-bar app spanning Codex, Claude Code and other AI coding tools with no separate login, reading local data. The choice is presentation and breadth. Use provider dashboards for authoritative billing detail; pick a notch viewer if you prefer that placement; and choose CodexBar when you want a menu-bar readout across several AI coding tools, always visible, drawn from data already on your machine, so you notice a limit approaching before you hit it.
MIT license and active development
CodexBar is MIT-licensed, so it is freely usable, and it is a native Swift macOS project with a widget extension, plugin integrations, and Homebrew distribution with an appcast for updates. The last push was on 2026-09-15, with frequent releases, a 0.60.x line, refining accuracy across providers such as keeping allowances tied to their own reset windows and improving spending totals, so it is very actively maintained. That cadence matters because the tools it reads and their usage schemes change, and a readout is only useful if it stays correct. Adopt it when you run AI coding tools on macOS and want their usage and spend visible in the menu bar without another login, install it with the Homebrew cask, confirm it shows your windows and totals, and treat the spending figures as usage estimates from local data rather than exact billing.
Editorial conclusion
Adopt CodexBar if you use AI coding tools like OpenAI Codex and Claude Code on macOS and want their usage windows and spending visible in the menu bar without a separate login. Do not expect it on other platforms or to manage your accounts, and treat its spending figures as estimates from local data rather than authoritative billing. Install it with brew install --cask codexbar, open the menu-bar item to confirm it reads your usage and totals, and rely on it for ambient awareness of approaching limits.
Frequently asked questions
What is CodexBar?
CodexBar is an MIT-licensed macOS menu-bar app that shows usage and spending stats for OpenAI Codex, Claude Code and other AI coding tools without a separate login, reading the data those tools already store locally.
How do I install it?
On macOS via Homebrew: brew install --cask codexbar, with a tap-based formula and other routes also listed. It then lives in the menu bar and reads local data, with no separate sign-in to configure.
Are its spending numbers exact?
Treat them as estimates. CodexBar derives usage and spending from data the tools keep locally, so it is a convenience readout for awareness rather than authoritative billing, and its accuracy depends on the tools' data formats.
Official sources
Add this badge to your README
If you maintain this project, the badge below links readers to this analysis and shows its maintenance status from the daily GitHub snapshot. Paste the markdown into your README; add ?metric=license or ?metric=stars to the image URL for a different field.
[](https://hysenlabs.com/projects/steipete-codexbar)
Community notes