TL;DR

Run /usage inside Claude Code. On a Pro, Max, Team or Enterprise plan it shows bars for your five-hour session limit and your weekly limit, with reset times, plus a breakdown of what used them. On API billing it shows the session's token count and an estimated cost.

Two other places show the same plan limits:

• claude.ai → Settings → Usage, which also covers your chat usage.
• A custom status line, which can show 5h: 42% 7d: 18% at the bottom of every session.

The problem

Claude Code doesn't show your remaining allowance by default while you work. The first sign of trouble is often the warning at 85%, or the limit message itself. By then a long task is half done.

There are also several numbers that look alike. A session cost in dollars, a token count beside the spinner, a percentage of the five-hour window, a percentage of the week. They measure different things, and only some of them matter for a subscription.

Option 1: Check Claude Code usage with /usage

/usage

What you see depends on how you're signed in. From Anthropic's costs page:

On a subscription (Pro, Max, Team, Enterprise):

• Plan usage bars for the session and weekly windows, with when each resets.
• A breakdown of recent usage attributed to skills, subagents, plugins and individual MCP servers, as percentages.
• Recommendations, one per row, for anything behind 10% or more of recent usage. A row like 25% Trim ~/code/api/CLAUDE.md · 49k tokens per request says how much usage the row applies to, not how much the change saves.
• Loops, for the heaviest /loop or scheduled tasks that ran recently.
• A usage-credits row if usage credits are on.

Press d or w to switch the breakdown between the last 24 hours and the last 7 days. The breakdown is computed from session history on this machine, so usage from other computers or claude.ai isn't in it. The plan bars come from Anthropic and do include everything.

When the footer shows a hint such as f to fix 2 of 3, pressing f has Claude Code act on the recommendations it can. That needs Claude Code v2.1.305 or later.

On API billing:

The Session block shows total cost, API and wall-clock duration, code changes, and tokens by model. Claude Code computes the dollar figure locally at list price. For the real bill, use the Usage page in the Claude Console. These totals reset when /clear starts a new session.

When /usage says rate limited

A message such as rate limited until 3:14am means Claude Code can't fetch your plan limits again until then. It isn't your usage limit. When the request fails, /usage shows the last bars it loaded on this machine within the past hour, with a Showing last-known usage note. Press r after the time shown for fresh data.

Option 2: Show Claude Code limits in the status line

For a number that's always visible, add the rate_limits fields to a custom status line. Claude Code passes JSON to your script on stdin. For Pro and Max subscribers, after the first API response, it includes:

• rate_limits.five_hour.used_percentage and rate_limits.five_hour.resets_at
• rate_limits.seven_day.used_percentage and rate_limits.seven_day.resets_at

used_percentage runs from 0 to 100. resets_at is Unix epoch seconds.

A minimal script, adapted from Anthropic's example:

#!/bin/bash
input=$(cat)

MODEL=$(echo "$input" | jq -r '.model.display_name')
FIVE_H=$(echo "$input" | jq -r '.rate_limits.five_hour.used_percentage // empty')
WEEK=$(echo "$input" | jq -r '.rate_limits.seven_day.used_percentage // empty')

LIMITS=""
[ -n "$FIVE_H" ] && LIMITS="5h: $(printf '%.0f' "$FIVE_H")%"
[ -n "$WEEK" ] && LIMITS="${LIMITS:+$LIMITS }7d: $(printf '%.0f' "$WEEK")%"

echo "[$MODEL] $LIMITS"

Save it as ~/.claude/statusline.sh, make it executable with chmod +x, and point statusLine in ~/.claude/settings.json at it:

{
  "statusLine": {
    "type": "command",
    "command": "~/.claude/statusline.sh"
  }
}

The // empty in the jq calls keeps the script quiet when the fields are missing, for example on API billing or before the first response.

Option 3: Check usage on claude.ai

Open Settings → Usage on claude.ai. It shows the current session and the weekly limit for the whole account, including chat. It's the place to look from a phone, or when you want to see what a chat session used.

Option 4: Check API spend in the Console

On pay-as-you-go, the Claude Console's Usage page is the authoritative record. A workspace called "Claude Code" is created the first time you authenticate Claude Code with a Console account, which keeps its spend separate from your other API use.

What the numbers mean

The spinner count, for example (10m 27s · ↓ 5.5k tokens), covers output only. Your prompt and the conversation history don't move it, so it says little about how much of the window a turn used.

Checking usage across agents and computers

/usage covers one Claude account on one machine's view. If you also run Codex or Gemini, or more than one Claude account, each has its own place to check.

If you want to see what's left on each of your accounts from your phone, Happier is an open-source app that lets you run and orchestrate Claude Code, Codex and other agents cross-device, from your phone, desktop or the web.

How its meters work:

• Three services have meters: a Claude subscription, a Codex subscription (openai-codex) and Gemini. API-key services and the other agents Happier runs have no quota meter.
• Gauges show what remains. 30% left · resets in 2h 10m means 30% of that window is still available.
• Stale data is labelled. Each account shows when it was last updated, marks old data as stale, and has Refresh quota now.
• The meters live with your connected accounts, in Settings → Connected services, and appear next to the account in a session.

Connected services and Connected services quotas are both experimental and off by default in Happier 0.2.17. Turn both on under Settings → Features. Setup is in Run Happier from your phone. Happier is on GitHub, you can download it here, and there's a Discord.

Troubleshooting

/usage shows a dollar cost on a Pro plan

That Session block is for API users. Anthropic's note says the session cost figure isn't relevant for billing on Pro and Max. If you expected plan bars and only see cost, check whether ANTHROPIC_API_KEY is set. If it is, Claude Code is billing the key, not the plan.

The status line shows nothing for limits

The rate_limits object appears only for claude.ai Pro and Max subscribers and only after the first API response. Send a message, then check again.

The breakdown doesn't match claude.ai

The breakdown uses session history on this machine. The plan bars and claude.ai include every device and chat.

FAQ

How do you check how much Claude Code usage is left?

Run /usage and read the session and weekly bars. They show used share and reset time.

Does /status show usage?

Anthropic's help article mentions /status for monitoring your remaining allocation. In current versions, /usage is the command the docs describe for plan bars and the breakdown.

Can you see Claude Code usage per project?

Not directly. The breakdown attributes usage to skills, subagents, plugins and MCP servers. For per-user numbers on a team, Anthropic points to analytics and OpenTelemetry export.

Why does a long session use so much?

Claude Code sends the full conversation with every request. A one-line question late in an all-day session still pays for the history. Running /clear between unrelated tasks keeps that history short.

Related guides

• When does Claude Code usage reset?
• What to do when you hit the Claude Code usage limit
• Is Claude Code free?
• How to use your Codex subscription on multiple computers