How to check your Claude Code usage

The direct answer

First: which kind of account are you?

“How much Claude Code have I got left?” has two very different answers depending on how you signed in. Getting this wrong is the most common reason the numbers look strange.

• A Claude subscription (Pro, Max, Team, seat-based Enterprise). You have usage limits: a session limit that resets every five hours and weekly limits. Anthropic says these limits are shared across Claude and Claude Code, and that IDE usage counts toward the same pool. The question is “what percent is used, and when does it reset?”
• An API key or Claude Console login. You pay per token. There’s no five-hour bar to watch; the question is “how much have I spent?”, and the answer is in the Console.
• A cloud provider (Amazon Bedrock, Google Cloud, Microsoft Foundry). Billed per token to your cloud account. Spend shows up in that provider’s billing console.

One trap: Anthropic’s help center warns that if ANTHROPIC_API_KEY is set in your environment, Claude Code uses that key instead of your subscription, and you get API charges. If you think you’re on Max but see dollar figures and no bars, check this first. For how the limits themselves work, see Claude Code usage limits, explained.

1. /usage in the terminal (start here)

Inside any interactive Claude Code session, type:

/usage

Claude Code’s command reference describes it as showing “session cost, plan usage limits, and activity stats.” It also runs immediately, even while Claude is in the middle of a response, so you don’t have to wait for a turn to finish. Here’s what’s on the screen:

Plan usage bars (subscribers)

Your current session and weekly limits, each with how far along you are and when it resets. On Max you may also see the separate weekly limit for Fable. These come from Anthropic’s servers, so they reflect your whole plan, not just this terminal.

If the usage request fails (most often because the usage endpoint is rate limited), /usage shows the last bars it loaded on this machine in the past 60 minutes with a Showing last-known usage note saying how old they are. Press r to retry.

The breakdown: what’s eating your allowance

On Pro, Max, Team and Enterprise, /usage also attributes recent usage to skills, subagents, plugins and individual MCP servers, flags behaviors such as long context or cache misses when one accounts for 10% or more of recent usage, and lists your heaviest /loop or scheduled tasks. Press d or w to switch between the last 24 hours and the last 7 days.

Important caveat from the docs: the breakdown is approximate and computed from local session history on this machine, so usage from other devices or claude.ai isn’t included. The bars are your whole plan; the breakdown is this Mac only.

The Session block (mostly for API users)

At the top sits a Session block: total cost, API and wall-clock duration, code changes and tokens by model. Anthropic says this block is intended for API users. The dollar figure is computed locally at list price, so on Pro or Max it isn’t your bill. It resets when /clear starts a new session.

Usage credits row

If you’ve turned on usage credits, /usage shows a row with this month’s credit spend against your monthly spend limit (or Unlimited if you haven’t set one).

/cost and /stats

Both still work, but they’re aliases. /cost opens /usage; /stats opens it on the Stats tab. Older blog posts that describe /cost as a separate cost-only command are out of date.

2. /status: are you on the account you think?

/status

The command reference says /status opens Settings on the Status tab, showing version, model, account and connectivity. It doesn’t show the usage bars. Anthropic’s Pro/Max help article still says to monitor your remaining allocation with /status, but in current builds the bars are under /usage.

Where /status earns its keep is the account line. If it shows an API key or a Console account when you expected your Max subscription, that explains why /usage shows dollars and no bars.

While you’re here: /context is about the context window (how full this conversation is), not your plan. A nearly full context means a long conversation, not that you’re near a usage limit.

3. claude.ai Settings → Usage (the whole picture)

Open claude.ai/settings/usage. Anthropic’s help center says Pro, Max, Team and seat-based Enterprise users see progress bars there:

• Current session: how much of the five-hour session limit you’ve used, plus the time remaining.
• Weekly limits: when your weekly limit resets, for all models and for Fable if your plan includes it.
• Usage credits: your credit settings, balance and this month’s spend.

This page is the best answer when you also chat on claude.ai or use several machines, because the limits are shared across claude.ai, Claude Code and Claude Desktop. It’s also where a one-time limit reset appears if Anthropic offers you one. Running /usage-credits from Claude Code opens this page in your browser on Pro and Max.

On a usage-based Enterprise plan there are no set limits; Anthropic says you can still track consumption under Settings → Usage.

4. The Claude Desktop app and VS Code

Desktop: the usage ring

In the Claude Desktop app’s Code tab, click the usage ring next to the model picker. Anthropic’s docs say it shows your current context window usage and your plan usage for the period. Context is per session; plan usage is shared across all your Claude Code surfaces.

VS Code: Account & usage

In the VS Code extension, /usage opens an Account & usage dialog. Signed in with a claude.ai plan, it shows usage bars for the current session and the week, each with how long until it resets, plus the same breakdown with a Day/Week toggle. With an API key or a cloud provider, it shows the session’s cost and tokens instead.

5. Keep it on screen: the status line

Typing /usage works, but only when you remember. Claude Code’s status line can show your limits all the time. For Pro and Max subscribers, the JSON Claude Code sends your status line script includes:

• rate_limits.five_hour.used_percentage and rate_limits.seven_day.used_percentage (0 to 100)
• rate_limits.five_hour.resets_at and rate_limits.seven_day.resets_at (Unix epoch seconds)

The docs note that rate_limits appears only for claude.ai Pro and Max subscribers (or behind a Claude apps gateway that sets a spend limit), only after the first API response in a session, and each window can be missing on its own. So your script has to cope with gaps. Run /statusline and describe what you want, or follow our step-by-step status line guide with a ready-made script. If you’d rather click than type, the free status line generator builds one for you.

The status line is live in the sense that it updates as you work: after each assistant message, when a rate-limit window reaches its reset time, and on a timer if you set refreshInterval. But it only exists inside an open Claude Code session.

6. API key users: the Claude Console

If you authenticate with a Console account or API key, you don’t have plan bars. Claude Code’s cost docs point you to:

• The Console usage page for authoritative spend. The /usage Session block is only a local estimate.
• Workspace spend limits. The first time you sign in with a Console account, a workspace called “Claude Code” is created automatically for tracking.
• The Console Claude Code dashboard for per-member spend and accepted lines.

Team and Enterprise admins get a spend report in org analytics on claude.ai. On cloud providers, spend lives in your cloud billing console, and Anthropic suggests OpenTelemetry for per-user numbers. Not sure whether a subscription would be cheaper than your API bill? Our API value calculator gives a rough comparison.

7. Local log tools like ccusage

ccusage is a popular open-source tool (not from Anthropic) that reads the usage data Claude Code saves on your machine and turns it into daily, weekly, monthly and per-session reports. Run it with:

npx ccusage@latest

It also has a blocks report for five-hour windows and a beta statusline command, and it can read Codex and other coding CLIs too.

What it’s good at: “which day was heavy?”, “which project used the most tokens?”, and rough API-equivalent cost. What it can’t do:

• It doesn’t know your plan’s limit. It counts tokens; Anthropic doesn’t publish limits as token counts, so it can’t tell you “62% of your weekly limit.”
• It only sees this machine. Chats on claude.ai, the Desktop app or another computer draw on the same limits but aren’t in your local logs.
• Costs are estimates. Its own docs say so. On a subscription they aren’t what you pay.

Use it for patterns. Use /usage or Settings → Usage for “how much is left.” If you want a report on how you work rather than how many tokens, Claude Code’s built-in /insights writes one to ~/.claude/usage-data/report.html (it spends some of your usage to make it).

8. Menu bar apps (always in view)

Most methods above share one weakness: you have to go and look, or have a Claude Code session open. A menu bar app keeps the number in the corner of your screen. Quotchi is ours, so take this with a pinch of sloth fur:

• It reads your Claude plan limits through Claude Code (2.1 or later), signed in with your Pro or Max subscription. It shows the five-hour and weekly windows and their reset times. API key and cloud-provider sign-ins have no plan limits to show.
• It shows Codex next to Claude, and Gemini app usage if you want it, so you can see which tool has room.
• Alerts at 80% (or your own levels), when a limit is reached, and when a fresh reading shows it’s back.
• A quotchi status terminal command and an opt-in status line segment.
• Free, no account, and it never reads your prompts.

Some model-specific limits may be missing from Quotchi, and the source is the same plan data /usage shows, so when in doubt, Settings → Usage is the referee.

Every method side by side

Troubleshooting: the numbers look wrong

/usage shows dollars but no bars

You’re probably signed in with an API key or Console account. Check /status, and check whether ANTHROPIC_API_KEY is set. Sign in with your subscription via /login if that’s what you meant to use.

The bars are higher than your terminal work explains

Limits are shared with claude.ai and Desktop, and with any other machine. The /usage breakdown only covers this machine, so it can look small next to the bars. Settings → Usage shows the combined picture.

“Showing last-known usage”

The live request failed, usually because the usage endpoint is rate limited, so Claude Code shows bars from up to 60 minutes ago. Press r to retry.

Your status line shows nothing for limits

rate_limits only appears after the first response in a session, only for Pro and Max, and a window disappears once its reset time passes. Send a message, then look again. More fixes in the status line guide.

ccusage and /usage disagree

They measure different things. ccusage counts local tokens and estimates cost; /usage shows your plan’s own percentage. Trust /usage for “how much is left.”

If the numbers say you’re (nearly) out

Anthropic says you’ll see warning messages as you approach a limit, and the limit message itself tells you which limit you hit and when it resets. From there: let Claude Code wait and continue automatically after a session reset (v2.1.234 or later, claude.ai sign-in), turn on usage credits with /usage-credits, use /model if only an Opus or Sonnet limit is blocking you, or switch tools. The full playbook is in Claude Code usage limit reached: what to do, and when does Claude Code reset? explains the timing. Got a reset time from an error message? The Claude reset time tool turns it into a countdown in your time zone. If this happens every week, Max vs Pro helps you decide whether to upgrade.

Common mistakes and myths

“/cost shows what I’m paying on Max.”

It opens /usage, and the Session cost there is a local list-price estimate. On Pro and Max, your usage is included in the subscription.

“/status shows my remaining usage.”

In current builds it shows version, model, account and connectivity. The bars are in /usage.

“Terminal usage is separate from claude.ai.”

On Pro and Max the limits are shared across Claude and Claude Code, IDE usage included.

“ccusage knows my limit.”

It knows your local tokens. Your limit percentage comes from Anthropic, via /usage or Settings → Usage.

“Context full means usage full.”

/context is about this conversation. Your plan limit is a different number.

What changed recently

Sources

Official pages (and one project README) we read on October 17, 2026. If anything here disagrees with them, they win.

• Claude Code docs: Commands
• Claude Code docs: Manage costs effectively
• Claude Code docs: Status line
• Claude Code docs: Desktop application
• Claude Code docs: VS Code
• Claude Help Center: Use Claude Code with your Pro or Max plan
• Claude Help Center: Usage limit best practices
• Claude Help Center: How do usage and length limits work?
• ccusage on GitHub and ccusage guide

FAQ