Usage and limits · Tenfold docs

Tenfold reads Claude Code’s own status line, so it knows your weekly limit and where your current pace lands by the reset. It also works out what your usage would cost at API rates, from Claude Code’s files on your Mac.

Where the figures come from

Claude Code passes a block of data to its status line every time it refreshes: the model, the context window, the session’s cost and, on a Claude plan, your rate limits. Tenfold wraps your status line so it can read that data, and yours keeps running as before. See What Tenfold changes.

You can turn this off in Settings, on the Integrations page, with the Status line switch. Without it, Tenfold no longer reads the model, context, cost and limits from the status line.

The weekly meter

The usage pill in the Workspace toolbar shows your weekly window: a meter, the percent used, and a countdown to the reset. It turns amber from 70% and rose from 85%.

Hover it for more:

The reset time, and the pace projection: where you’ll be by the reset if you keep going at the current rate, drawn as faded blocks. It reads like “~19% by reset”. If you’re on course to run out first, it says when, like “out in ~42m”. At 100% it says “limit reached”.
Per-model weekly windows, when your plan has them. These come from Claude Code’s last usage check, so the card says how old they are and has a button that runs /usage to renew them.
What you’ve spent today.

Click the pill to open the usage screen.

A few things to know:

Claude Code only reports Claude Code usage limits for Pro and Max plans. With an API key, there’s no weekly window to show.
The meter appears after Claude’s first reply in a session.
Tenfold waits until 5% of the week has passed before it projects your pace. Earlier than that, the figure is mostly noise.

The usage screen

Open it with ⌘U, from Usage History in the Tenfold menu, or by clicking the usage pill. It shows:

This month so far, with a pace bar: what you’ve spent, where you’re on course to end the month, and where last month ended.
Today, all time, and what caching saved you this month.
A spend calendar, one square per day for the last few months. Click a day to see what it cost.
Where the money goes: the cost split by token type (input, output, cache write, cache read).
Models: cost and tokens per model.

You can copy a text summary or a share card image from the top of the screen.

API-equivalent cost

Costs are shown as the API equivalent, at standard rates. They’re not a bill. Your agents run on your own Claude plan or API key, exactly as they would in a terminal, and Tenfold doesn’t add to your Claude usage. The figures help you see how much your plan is worth.

Claude’s total cost comes from Claude Code’s own figures. The split by model and token type is an estimate from token counts, scaled to add up to that total. Subscription charges and tool or service fees aren’t included.