Codex
Codex usage, without /status.
Codex meters your ChatGPT plan per window, and which windows you get depends on the plan: a 5 hour and a weekly window on Plus, weekly only on the Pro tiers. The CLI reports them behind /status.
The built-in way. Ask the CLI.
Run /status in the Codex CLI.
It lists each window with the percentage used and the reset time. Then you switch back to whatever you were doing and forget the number.
What Notchlet shows. One ring per window.
- 5h
- A rolling window, on the plans that have one.
- Weekly
- Reset once a week.
- Per plan
- Notchlet shows whatever the endpoint returns for your plan, with the labels and durations it sends. There is no plan table to go stale.
Hover the notch and each window is a ring with the percentage left, the reset time and a verdict on your burn rate. Click a ring for history: cost and tokens per day, and every model you used.
How it signs in. With the login you already have.
Notchlet reads the tokens the Codex CLI stores in ~/.codex/auth.json and calls the endpoint behind /status. It never refreshes them: they last ten days, the CLI renews them on use, and OpenAI rejects a reused refresh token, so a refresh behind the CLI's back could sign it out.
Questions. Short answers.
- How do I check my Codex usage?
- Run /status in the Codex CLI, or hover the notch with Notchlet installed.
- Does it work with a ChatGPT Plus plan?
- Yes. Any plan the Codex CLI signs into. The windows come from the endpoint per plan.
- Does Notchlet need my OpenAI API key?
- No. It uses the Codex CLI's login.