Claude Code Usage: How to Check It and What Uses It

Check your Claude Code usage with /usage, /status and /context, see what is using it, track API spend in the Console, and the habits that burn usage fastest.

Josh Pocock7 min read

Claude Code usage cover

AI Coding

Claude Code usage

Stride AI Academy

To check your Claude Code usage, run /usage in a session. On a Pro, Max, Ultra, Team or Enterprise plan it shows your plan usage bars and a breakdown of what used them, such as subagents, skills, plugins and MCP servers. On an API key it shows the session's token count and estimated cost. /cost and /stats open the same screen.

This is a how-to page: the commands that show usage, what each one tells you, where API users see spend, an open-source tracker, and the habits that use the most. For how the five-hour and weekly limits themselves work, see our guide to Claude usage limits. Everything here was checked against Claude Code's costs, commands and status line docs on November 22, 2026.

The commands that show Claude Code usage

Command What it shows
/usage Session cost, plan usage limits, activity stats and a usage breakdown
/cost Alias for /usage
/stats Alias for /usage, opened on the Stats tab
/status Version, model, account and connectivity
/context A grid of what's filling the context window right now
/insights An HTML report on how you work across recent sessions
/usage-credits Your usage-credits settings, or a request to your admin

The commands reference says /status and /usage run straight away, even while Claude is responding.

How to check usage with /usage

Run it at the prompt:

/usage

What you see depends on how you sign in.

On a subscription plan, the costs docs say you get plan usage bars, activity stats and a breakdown of what counts against your limits:

  • Attribution: recent usage split between skills, subagents, plugins and individual MCP servers, each as a percentage.
  • Recommendations: suggested changes, one per row, each starting with the share of your usage it applies to. The docs give 25% Trim ~/code/api/CLAUDE.md · 49k tokens per request as an example.
  • Loops: the heaviest /loop and other scheduled tasks that ran recently, with their token totals.

Press d or w to switch between the last 24 hours and the last 7 days. On v2.1.305 or later, if the footer shows a hint like f to fix 2 of 3, press f and Claude Code applies the recommendations it can, showing you each change first.

One limit to know: the breakdown comes from session history on that machine. Usage from claude.ai or your other computers isn't in it. For your full plan picture, use Settings > Usage on claude.ai.

On an API key, the Session block shows the session's token usage and an estimated dollar cost:

Total cost:            $0.55
Total duration (API):  6m 20s
Total duration (wall): 6h 33m 10s
Total code changes:    0 lines added, 0 lines removed

The docs say this figure is computed locally from token counts at list price, so treat it as an estimate. It resets to $0 when /clear starts a new session. Subscribers can ignore it, because their usage is included in the plan.

Check which account you're using with /status

/status

This shows your version, model and account. It's worth running when usage looks wrong. The setup docs explain the usual trap: if ANTHROPIC_API_KEY is set in your environment and you approve it when asked, Claude Code uses that key instead of signing you in to your plan.

See what's filling your context with /context

/context

Every request sends your context, so a heavy context means heavy usage. /context shows a colored grid of what's taking space, with suggestions for context-heavy tools and oversized memory files. Pass all to expand the per-item breakdown. Our guide to Claude Code context management covers what to cut.

Watch usage all the time in your status line

To see your allowance without running a command, put it in a custom status line. The status line docs list these fields for Pro, Max and Ultra subscribers:

  • rate_limits.five_hour.used_percentage and rate_limits.seven_day.used_percentage, from 0 to 100
  • rate_limits.five_hour.resets_at and rate_limits.seven_day.resets_at, as Unix epoch seconds

They appear only after the session's first API response, and each window can be missing, so read them with a fallback:

Terminal
jq -r '.rate_limits.five_hour.used_percentage // empty'

For API users, cost.total_cost_usd gives the estimated session cost. Run /statusline and describe what you want, and Claude Code writes the script for you.

Get a report on your habits with /insights

/insights

This analyzes your recent sessions on that machine and writes an HTML report on what you work on, where things went wrong and features to try. The docs say it saves to ~/.claude/usage-data/report.html. It runs through your normal account, so the report itself uses some of your usage.

Check Claude Code usage as an API user

If you sign in with a Claude Console account, you're billed per token, and the costs docs point you to the Console:

The first time a member signs in, Claude Code creates a workspace called "Claude Code" in your Console organization, which you can give its own rate limit. Our Claude API pricing guide covers the per-token rates, and the Claude Code cost calculator estimates a monthly bill.

On Team and Enterprise

Admins see estimated spend per user and per model in the spend report under org analytics, with CSV export. The docs note that usage inside the seat allowance isn't metered in dollars; the report covers usage-credit spend once credits are on. Members can still run /usage for their own bars.

Track usage with an open-source tool

The costs docs also mention OpenTelemetry export for teams that want per-user metrics in their own monitoring stack. For a quick personal report, a popular open-source option is ccusage, which reads the local usage data coding agents save and turns it into reports. Its repo is github.com/ccusage/ccusage; on November 22, 2026 the GitHub API showed it as MIT-licensed with 19,540 stars and a push that day. It isn't made by Anthropic.

The README runs it without installing:

Terminal
npx ccusage@latest

Its usage examples run the reports with Bun's bunx:

Terminal
bunx ccusage daily
bunx ccusage monthly
bunx ccusage session
bunx ccusage blocks

blocks groups Claude Code usage into five-hour windows. Like /usage, it only sees sessions on the machine you run it on.

What uses the most Claude Code usage

The costs docs list why a long session can use far more than your activity suggests:

  • Long context. Your full conversation goes out with every request, and each batch of tool results adds another request.
  • Cache misses. Your first message after a long break can miss the prompt cache and reprocess everything.
  • Scheduled tasks. A /loop fires on its interval even while you're idle.
  • Subagents and workflows. Each one sends its own requests on top of yours.
  • Agent teammates. Each keeps using tokens until it exits.
  • Compaction. /compact reads the whole conversation it summarizes.

The fixes from the same page:

  1. Run /clear between unrelated tasks. It costs nothing.
  2. Match the model to the job. The docs say Sonnet handles most coding tasks and costs less than Opus. Switch with /model.
  3. Lower the effort with /effort for simple tasks.
  4. Keep CLAUDE.md under 200 lines and move occasional workflows into skills, which load only when used.
  5. Turn off MCP servers you aren't using with /mcp.
  6. Write specific prompts. "Add input validation to the login function in auth.ts" reads far fewer files than "improve this codebase."

If you keep running out even with good habits, our guides to Claude credits and how much Claude Code costs compare your options.

Learn to build with Claude Code

The Stride AI Coding course teaches Claude Code and Codex for building client tools, written for non-programmers, with weekly live calls to review your builds. Many of the free guides cover Claude Code setups, and the Claude Code community is where members compare theirs.

FAQ

How do I check my Claude Code usage?

Run /usage in a Claude Code session. On a subscription it shows your plan usage bars and a breakdown of what used them. On an API key it shows the session's tokens and estimated cost. /cost and /stats are aliases.

Is /cost still a Claude Code command?

Yes, but the commands reference lists it as an alias for /usage. Typing either opens the same screen.

Why doesn't /usage match what claude.ai shows?

The breakdown in /usage is computed from session history on that machine, so it leaves out claude.ai and your other devices. Settings > Usage on claude.ai shows your whole plan.

How can I see Claude Code usage all the time?

Add a custom status line. Pro, Max and Ultra subscribers get rate_limits.five_hour.used_percentage and rate_limits.seven_day.used_percentage fields. Run /statusline to have Claude Code set it up.

Where do API users see Claude Code spend?

In the Claude Console. The Cost page shows spend, the Claude Code dashboard shows spend per member, and workspace spend limits cap it. The Claude Code Analytics API returns the same data in code.

What uses the most Claude Code usage?

Long sessions that are never cleared, subagents and agent teams, scheduled loops, and cache misses after long breaks. Run /clear between tasks and check the /usage breakdown for your biggest share.

Build it with us in Stride AI Academy

Four courses, weekly live calls and a community of people building and selling AI. Or start with 320+ free guides.

Keep reading

Free AI Coding guides