Type /usage in Claude Code. That shows how much of your plan you have used. To
see which session used it, you need the per-session numbers, which is the part most tools leave out.
1. The /usage command
Run /usage inside any session (/cost does the same thing). It shows:
- Your plan limits, as bars, if you are on a Pro, Max, Ultra, Team or Enterprise plan.
- A breakdown of what counted against them: the share of recent usage that went to skills,
subagents, plugins and individual MCP servers, with suggested fixes. Press
dorwto switch between the last 24 hours and the last 7 days. - A session block with token counts for the current session. That part is meant for people paying per token through the API.
Two limits worth knowing: the breakdown is worked out from session history on the machine you are on, so work done on another computer is not included, and it groups usage by category, not by session.
2. The /context command
Run /context to see what is filling the current session's context window right now: instructions,
tool lists, skills, and the conversation itself. Run it at the very start of a fresh session to see how big the
session is before you have done anything.
3. A status line
Claude Code can show a line at the bottom of the terminal that updates as you work, including how full the context window is and how much of your 5-hour and weekly limits you have used. See how to set up a status line that shows cost and context.
4. The log files
Every session is written to ~/.claude/projects/, with the token counts for each reply. Adding
those up per file gives you usage per session and per day. It is the only one of the four that covers every session
you have ever run on the machine. Where the logs are and how to read
them.
Which session used it?
This is the question the built-in tools answer least directly. On the author's Mac, one session of several was usually far ahead of the rest, and three numbers explained why:
| Number | What it tells you |
|---|---|
| Size at the first turn | How much is loaded before any work. Ordinary sessions started at about 126K. |
| Size per turn now | How much is re-read on every reply. One session reached 300K within an hour. |
| Share by subagents | Whether helpers did the spending. On the heaviest day they made 61% of turns. |
The session figures on this page come from real logs on the author's Mac, read with Cost Doctor. Your numbers will differ; the pattern usually does not.
Cost Doctor shows that list in your menu bar: today's sessions ranked by tokens, each with its start, its size now and one line of advice. It reads the same log files, includes subagents, and covers Codex too.
See it on your own sessions
Cost Doctor is a Mac menu bar app that reads the logs Claude Code and Codex already write. It names the session that is using your tokens, says why, and turns your screen red when a check fails while an agent sits waiting. Nothing leaves your Mac.
Download the free trialHow it works14 days free, no account. Then $19 once, plus tax where it applies. For Apple silicon and Intel Macs on macOS 13 or later.