If you want to know how to check Codex usage, start with /status in the Codex CLI: it shows your rate limits right where you work. For reset times open the Codex usage dashboard on chatgpt.com, and for history, credits and alerts across days, run UsageAtlas, a free, open-source desktop tracker that reads the Codex CLI you are already signed in to.
Each of the three answers a different question. The rest of this guide walks through them in the order most people need them.
Step 1: run /status in the CLI
OpenAI's slash command reference describes /status as the command that shows "the chat ID, context usage, and rate limits." Type it at the Codex prompt and you get the current picture for your account.
Use it when the only question is whether to send the next prompt. It is quick, it needs no extra setup, and it lives inside the session. What it does not give you is a record. You cannot see last Tuesday, a weekly trend or an estimate of what the work cost.
One detail trips people up. "Context usage" in that output is how full the model's working memory is in this chat. It is not your plan allowance. A chat can be close to its context limit while your account has plenty of room, and the other way round.
Step 2: open the usage dashboard
OpenAI points users to the Codex usage dashboard at chatgpt.com/codex/settings/usage for current limits and reset times. If you signed in to Codex with ChatGPT, the /usage command opens usage and billing settings for you.
This is the page to trust for the official numbers. OpenAI says its per-plan figures are estimates, not fixed message limits, because the model, the size of the task and whether it runs locally or in the cloud all change how much each message uses. So rather than memorize a number, check the dashboard.
Step 3: track it over time on the desktop
The CLI and the web page both show now. If you want the trail, UsageAtlas fills that gap:
- Download the app for Windows, macOS or Linux.
- Let it use the signed-in Codex CLI on the same machine. It talks to the CLI's local app-server API, not to a UsageAtlas cloud account.
- Open Today for session and weekly windows, remaining capacity, plan and credits where Codex reports them.
- Open History for daily totals, tokens, models and estimated cost.
- Set an alert threshold so you get one desktop warning per window before you run dry.
No account is needed for local tracking, and the app does not read your prompts or responses. An optional free account saves your history so you can restore it on another computer.
Which check to use when
| Question | Best place |
|---|---|
| Can I send another prompt right now? | /status in the CLI |
| When does my limit reset? | Codex usage dashboard, or UsageAtlas |
| How much did I use this week? | UsageAtlas History |
| What did a day cost, roughly? | UsageAtlas (estimate) |
| What am I billed? | Your ChatGPT billing settings |
Cost figures in UsageAtlas are estimates. For billing, OpenAI's own settings remain the record.
When the built-in checks are enough
If you use Codex lightly and only ever ask "am I close?", /status and the dashboard cover it, and you do not need another app. A tracker earns its place once Codex is a daily tool, when you hit limits mid-task, or when you also run Claude Code, Cursor or Copilot and want them in one view. UsageAtlas covers those tools in the same window, which is the main reason to pick it over checking each vendor separately.
The project's source is on GitHub under the MIT license.