How to check your Claude Code usage
Type /usage in Claude Code, or open Settings › Usage on claude.ai, to see your weekly limit and, on Pro and Max, your current 5-hour session. Both read the same pool.
Contents
Type /usage in Claude Code. On a Pro, Max, Team or seat-based Enterprise plan it shows your plan's usage bars, activity stats and what's been eating your limits. Outside the terminal, open Settings › Usage on claude.ai: it has progress bars for your current 5-hour session and your weekly limit. A verified Ultra account has no 5-hour limit, so it shows no "Current session" bar (Anthropic's Ultra page). Both read the same pool.
In Claude Code: /usage#
Anthropic's command reference describes /usage as showing "session cost, plan usage limits, and activity stats". /cost and /stats are now aliases for it, so old guides that tell you to run /cost still work and land you on the same screen (/stats opens it on the Stats tab). /usage runs immediately, even while Claude is in the middle of a turn; you don't have to wait for the turn to finish.
The Session block#
At the top, the cost docs list total cost, API time, wall time, lines changed, and tokens per model, split into input, output, cache reads and cache writes. Anthropic is clear that the dollar figure is for API users. If you're on Pro or Max, your usage is included in the plan, so "the session cost figure isn't relevant for billing purposes." It's still a handy yardstick for how heavy a session was.
Since v2.1.251 there's also a Prompt cache (main) line: how much of your input came from cache, how many misses, and whether the cache is warm right now. If the misses climb, the usual causes are in why Claude Code uses so many tokens.
The Session totals reset when /clear starts a new session.
Plan usage bars and the breakdown#
On a subscription you also get your plan's usage bars and a breakdown of what counts against them. Attribution splits recent usage across skills, subagents, plugins and individual MCP servers, as percentages. Recommendations come one per row, each covering at least 10% of recent usage, such as trimming a heavy CLAUDE.md; on v2.1.305 or later, an f to fix hint lets Claude Code apply the ones it can. Loops lists the heaviest /loop and other scheduled tasks that ran recently.
Press d or w to switch between the last 24 hours and the last 7 days. Those figures come from this machine's local session history, so work on another computer or on claude.ai isn't in the breakdown.
For more than the last 7 days, our free Claude Code usage dashboard reads the same local transcripts with one command and draws every day they cover: models, tokens, API-equivalent cost and busiest hours, on a private page.
When it says "rate limited"#
If the request for your plan limits gets rate limited, /usage shows the last bars it loaded in the past 60 minutes with a Showing last-known usage note. A message like rate limited until 3:14am means Claude Code can't fetch your limits until then. It doesn't mean your work is limited. Press r to retry after that time.
/status and /context#
/status shows your version, model, account and connectivity. Anthropic's help article on Pro and Max suggests it for keeping an eye on your allocation, and it's the quickest way to confirm which account a terminal is actually using.
/context shows what's filling the current context window. That's not your plan limit, but a big context is the usual reason a limit drains fast.
On claude.ai: Settings › Usage#
Anthropic's usage limit best practices article says that on Pro, Max, Team and seat-based Enterprise plans you can go to Settings › Usage to see progress bars for your current session and your weekly limits. The session bar shows how much of your plan's five-hour session limit you've used and how long is left in that session. The weekly one shows when your plan's weekly limit resets, for all models.
This page is also where Pro and Max users manage usage credits. Running /usage-credits in Claude Code opens it for you.
What the numbers mean#
Anthropic's help pages, read on 22 October 2026, said paid plans had two clocks: "a session limit that resets every five hours" and "weekly limits that reset once every week." Since 30 October 2026 that holds for Pro and Max, while a verified Ultra has only its weekly limits. The 5-hour window is the one you feel in an afternoon. The weekly one is the one you feel on Thursday (5-hour limit, weekly limit).
It's one pool. Usage across Claude's products, including Claude Code, counts toward the same limit, so a long chat on claude.ai drains the same bars as your agents.
Some limits are per model. Anthropic's Claude Code article says "Max and Ultra plans also have a separate weekly limit for Fable." If you see a model-specific limit message, switching to a model outside that family keeps you working. A session or weekly limit covers every model, though, so /model won't help.
Anthropic doesn't publish a fixed number of messages per window, and neither will we. How far a window goes depends on the model, the size of your context and how much your agents do per turn.
If you pay per token (API keys)#
With an API key, there's no plan bar to fill. You're billed per token, so the number that matters is cost.
- The Session block in
/usageis computed locally from token counts at list price. Anthropic calls it an estimate and points to the Usage page in the Claude Console for authoritative billing. - You can put cost in your status line so it's always visible.
- On Team and Enterprise plans, admins see spend in the org's analytics and can export it per user.
| What you want to know | Where to look |
|---|---|
| How much of my 5-hour and weekly limit is left | /usage in Claude Code, or Settings › Usage on claude.ai |
| What's eating my limit | /usage breakdown (subscription plans) |
| Which account this terminal uses | /status |
| What's filling the context window | /context |
| What I'll be billed on an API key | Claude Console › Usage (/usage is an estimate) |
Watching several accounts at once#
Everything above shows one account: whichever one that terminal or browser is signed in to. If you have a personal Pro and a work Max, checking both means two sign-ins, or two configuration directories, and keeping the reset times in your head.
That's the routine we got tired of, so Shuriken shows them together. In Settings › Sources, every account you added has two bars side by side: its 5-hour window, marked S, and its weekly limit, marked W, each with the share left written inside. Point at a bar and it tells you when that limit resets. The bars stay grey while there's room, turn amber from 75% used and red from 90%, the same rule everywhere they appear. The top bar shows how much you have left this week across your accounts.
One device reads each account and shares the reading with your other devices, sealed with your own key. Your laptop and your server agree on how full an account is, and they don't both ask. If a read is refused for a while, you keep the last good number, marked with its age. Codex and other providers' plans have readers too, wherever the provider reports usage.
The readings also drive source switching, which moves your chats to the account with the most room when the one in use reaches your switch point (99% by default). We wrote up the feature in every Claude account's limits on one screen.
For spend rather than limits, the Analytics page, under Home, shows what your Claude Code and Codex turns would cost at API prices, by account, project and model, over days or months. And Shuriken 5.1 adds live cost: every chat shows what its current turn costs while it runs, beside the turn's time, for Claude Code, Codex and other agents that report usage as they go.
Only add accounts you own and are entitled to use. If juggling /usage across accounts is part of your day, download Shuriken and add them as sources.
Try the free tool: Claude limit reset tracker
Sources
- What is the Ultra plan? (Claude Help Center) (accessed 30 October 2026)
- Commands (Claude Code docs) (accessed 22 October 2026)
- Manage costs effectively (Claude Code docs) (accessed 22 October 2026)
- Usage limit best practices (Claude Help Center) (accessed 22 October 2026)
- How do usage and length limits work? (Claude Help Center) (accessed 22 October 2026)
- Use Claude Code with your Pro or Max plan (Claude Help Center) (accessed 30 October 2026)
Frequently asked questions
What's the difference between /usage and /cost?
There isn't one any more. In current Claude Code, /cost and /stats are aliases for /usage; /stats opens it on the Stats tab.
Can I see usage for all my Claude accounts at once?
Not in Claude Code or on claude.ai: each shows the account you're signed in to. Shuriken shows every account you added as a source side by side, with its 5-hour and weekly bars.
Can I see my usage history?
In /usage, press d or w to switch the breakdown between the last 24 hours and the last 7 days. It's computed from this machine's local history, so other devices and claude.ai aren't included.
Does chatting on claude.ai count against my Claude Code limits?
Yes. Anthropic says usage across Claude product surfaces, including Claude Code, counts toward the same limits.
Why does /usage say I'm rate limited?
The request for your plan limits was rate limited, not your work. /usage then shows the last bars it loaded in the past hour, and you can press r to retry later.