Codex Usage Tracker and Dashboard

Connect Codex

Start your board Sign in at board.contextbolt.com with your email. The Agents template already has the AI usage widget on it. On another board, add AI usage from the gallery yourself, since it reads your own computer.
Link your computer In Connect, open Your AI, then Codex, and copy the line that links your computer. Run it once in a terminal. The small runner starts with your computer from then on.
Use Codex as usual The rings fill the next time Codex runs on that computer. They empty as you use your plan and refill when each window resets. No command to type.

Your limits are read on your computer, and only the percentages and reset times reach the board. How the runner works

What to ask your agent

“Arrange this board so my AI usage sits top left, with Needs you beside it” Your agent moves the widgets into place, so your Codex limits are the first thing you see when the board opens.

“Add a note by my usage: big refactors go to Claude Code when Codex is under 20%” A sticky note beside the rings. The rule you set for yourself stays in sight at the moment it matters.

“Post an update to my board when this migration is done” Add the board to Codex as well and it posts a line to Board activity when it finishes, so you can walk away from the terminal.

“Add 'review the API session's plan' to my to-do list” Your agent finds your to-do list by itself. If you have two, it asks which one.

What you see from Codex

See what is left of your Codex five-hour window, as a ring
See what is left of your week, with the day and time it resets
Watch Codex and Claude Code side by side in one widget
Follow every Codex session on your computers, working, done or waiting
Spot the session waiting on your OK, in Needs you
Add usage from more than one computer, or a server, with the same one line
Keep it next to your revenue, traffic and to-dos, on a board you arrange

Codex on Context Board: the full guide

You find out where your Codex limit is the worst way. Halfway through a task, when Codex stops and tells you to try again later.

Codex will tell you where you stand if you ask. Type /status in the CLI, or open the usage page at chatgpt.com/codex/settings/usage. The trouble is remembering to ask while you are deep in the work.

What the board shows

The AI usage widget draws your Codex plan as rings. One for the current five-hour window, one for the week. Each ring empties as you use it, turns amber under half and red under a fifth, and says when it comes back, like “resets in 2h 14m” or “resets Thu 2:00 PM”.

It draws the windows Codex reports for your plan. On Plus that is a five-hour window and a week. OpenAI says Pro plans currently have no five-hour limit, so there the week is the one to watch.

Claude Code sits beside it in the same widget, read the same way. So when Codex is down to 15% of its week and Claude Code still has 70%, you know which one gets the next big task.

Where the numbers come from

Codex already knows your limits. Each time it runs, it hears from OpenAI how much of the window and the week you have used, and keeps that count in its own session files on your computer.

The board’s runner, a small program you set up once with one line, reads the newest of those counts and sends your board the percentages and reset times. Nothing else. Not your prompts, not your code, not what Codex wrote back. The widget says when Codex last checked, so a number is never older than it looks.

One thing to know. The count moves when Codex runs on a linked computer. OpenAI says local and cloud use share one allowance, so after an hour of Codex in the cloud, the rings catch up the next time Codex runs on your machine.

Codex usage at a glance

/status answers inside one terminal, when you remember to type it. /statusline can keep your limits in the CLI’s footer, which helps, as long as you are looking at that terminal. A board answers when you glance at it, whichever window you are in.

Every new account gets an Agents board with the usage widget already on it, beside your Codex and Claude Code sessions, the one waiting for your OK and today’s commits. Put your revenue next to it, or anything else you check in a day.

If you run Codex on more than one computer, or on a server, link each one with the same line. Usage and sessions come from all of them.

Frequently asked questions

Something else on your mind? Get in touch.

How do I check my Codex usage?

Inside Codex, type /status in an active CLI session to see what is left of your limits. On the web, your usage dashboard is at chatgpt.com/codex/settings/usage. OpenAI's Codex pricing page points to both. Context Board shows the same limits all the time, with reset times, without you asking.

How does Context Board know my Codex usage?

Each time Codex runs, it saves what OpenAI told it about your limits in its own session files. The runner, a small program you set up once with one line from Connect, reads the newest of those counts and sends the board the percentages and reset times. Nothing else leaves your computer.

What if my plan has no five-hour limit?

The widget draws the windows Codex reports for your plan. OpenAI says Pro plans currently have no five-hour limit, so on Pro the week is the ring to watch. On Plus you see both.

Does it work with Claude Code as well?

Yes. The same widget shows Claude Code's five-hour session and week beside Codex, read the same way, plus a model's own weekly limit when your plan has one. See Claude Code usage on Context Board.

What does it cost?

Context Board is free for 3 days, then $6 a month. One plan with every connection, widget and theme. The runner and the AI usage widget are part of it.

Does it change anything in Codex, or send my code anywhere?

No. Reading your limits changes nothing in Codex and spends none of your plan. From your sessions, only names, states and percentages leave your computer, never your prompts, your code or what Codex wrote. If you run the board's chat or agent jobs on Codex, those use your plan like any other Codex session.

Go deeper

Keep reading

All integrations