17 / AI USAGE · 5 min read
How to Check Your Codex Usage and Limits
Start with what the Codex CLI itself reports, then see how to keep the numbers in view without typing a command.
In the Codex CLI, type /usage to see your account’s token activity; /usage daily, /usage weekly and /usage cumulative open a view directly. To keep your rate limits in sight, type /statusline and switch on the rate-limits item in the footer. Tegaro adds a menu-bar card with each window’s percentage left and its reset time.
Codex usage is limited by windows: a short rolling window and a longer weekly one, depending on your plan. Running into one mid-task is the usual way people find out, so it helps to know where to look first.
1. Type /usage in Codex
OpenAI’s slash-command reference says /usage opens a usage menu where you can show token activity or redeem an available earned reset. To open token activity directly, add daily, weekly or cumulative. It needs a session with Codex account sign-in; otherwise Codex asks you to sign in.
2. Pin the limits in the footer with /statusline
Type /statusline to open a picker. Among the items you can switch on and reorder is rate limits, alongside the model, context, git branch and token counters. The choice is saved in config.toml, so the footer keeps showing it in later sessions.
3. What /status shows
In the Codex CLI, /status is a session check: it prints the active model, approval policy, writable roots and current token usage. Use it to confirm where Codex is operating, not to read your plan’s limits.
4. Read the windows correctly
A plan can report a short rolling window, a weekly window, or only one of them. ChatGPT Pro, at the time of writing, reports a weekly window only. Some models also have their own separate limit, which the vendor names. If your plan has Codex credits, a used-up window can keep going on them, which is why “limit reached” and “out of Codex” are not always the same moment.
5. Mind the reset time
Percent left means little without the reset. A window at 90 % used that resets in ten minutes is a different situation from one that resets in six days. Always read the two together.
Where Tegaro fits
Tegaro AI usage puts Codex and Claude limits in the menu bar so you do not have to type a command.
- Each window shows “42 % left” and “resets in 2 h 10 min”, with the bar colour changing as it is used up: amber from 90 % used, red at 100 %.
- It reads the sign-in Codex already saved on your Mac and asks OpenAI for the limits with it. Provider credentials are not sent to Tegaro. If the answer fails, it falls back to the newest Codex session log and marks the card “as of” a time.
- Model-specific limits and Codex credits show on the status line, for example “Week used up” and “on credits” when your credits are covering it.
- A quiet Notification Center notice arrives when a window passes 80 % (once per window) and when it resets. No permission prompt appears.
- “Task finished” tells you when a Codex task that ran a minute or longer ends, so you can work elsewhere meanwhile. You choose the minimum length: 30 seconds, 1, 2 or 5 minutes (1 minute by default).
AI usage stays free after the 14-day trial and can be switched off at any time. It shows what the provider reports; it cannot raise a limit.
Windows, credits and plan rules are set by OpenAI and change over time. Trust the numbers the tools report today over any figure quoted in an article.
Frequently asked questions
How do I see my Codex usage limits?
In the Codex CLI, type /usage for your account’s token activity, and add the rate-limits item to the footer with /statusline to keep the limits visible.
What does /status show in the Codex CLI?
The active model, approval policy, writable roots and current token usage. It checks the session setup; it is not the command for plan limits.
Why does my plan show only a weekly Codex window?
Plans differ. ChatGPT Pro has, at the time of writing, reported only a weekly window. Tegaro then shows that there is no 5-hour limit on the plan.
Does Tegaro send my Codex login to its servers?
No. It uses the sign-in already on your Mac to ask OpenAI for the limits, and provider credentials are not sent to Tegaro.
Can I get a warning before I run out of Codex usage?
Yes. Tegaro posts a quiet notice when a Codex or Claude window passes 80 % and again when it resets. The switch is AI usage, then Notifications, then Limit alerts.