Claude Usage icon: a blue outer ring around an orange inner ring

Claude Usage

Your Claude limits, in the menu bar.

Two rings: your 5‑hour session inside, your weekly limit outside, each filling as you use it. A tick on each ring shows how far through the window you are, and turns red when you're on pace to hit the limit before it resets — so you find out now, not mid‑task.

Download coming soon

Free. For Apple silicon Macs with macOS 14 Sonoma or later.

Reads your limits from claude.ai with your own claude.ai sign‑in, and checks Claude's status page for outages. It connects nowhere else — not even to us.

The Claude Usage popover on the Usage tab: two concentric rings showing session usage at 1% and weekly usage at 8%, reset times, on-pace estimates of about 4% for the session and 11% for the week, a by-model share bar, claude.ai's weekly model limits, and extra usage of $0.00 of $20.00.
Session and weekly limits, reset times, and where you're headed.
The Models tab: this week's Claude Code usage by model, about $138 at API prices across 1,881 requests, with token and cached-read counts for Opus 5.5 and Haiku 4.5 and a daily bar chart.
Claude Code usage by model, for the session or the week.
The Settings tab: signed in to a Claude Max account, notifications at 80% warning and 95% critical and when the session and weekly limits reset, a two-minute refresh, launch at login, and the option to write usage for a Claude Code statusline.
Thresholds, reset alerts, and the statusline switch.

Pace, not just percent.

40% means nothing without knowing how much of the window is left. Each ring's tick marks the time gone: green when you're on track, orange when it's close, red when you'll run out before the reset. The rings themselves turn orange and then red as you near the limit.

See where it went.

The Models tab breaks your Claude Code usage down by model for the session or the week, with tokens, requests, and what it would have cost at API prices. It reads Claude Code's own logs on your Mac, and only reads them.

Warned, then told it's back.

A notification when you cross your warning and critical thresholds, and another when the session or weekly limit resets — so you can stop rationing the moment you don't have to. The History tab keeps up to 90 days, so you can see how your weeks usually go.

How it works

  1. Sign in to claude.ai.

    In a login window inside the app, the same way you would in a browser — or paste a session key if you'd rather. The session lives only in the app's own Keychain item. No API key needed.

  2. Glance at the rings.

    They're the same numbers claude.ai's own usage page shows, refreshed every couple of minutes. Click for the details.

  3. Optionally, put it in Claude Code too.

    Turn on Write usage for a statusline and the bundled statusline script shows your model, folder, and context alongside session and weekly bars, under Claude Code's prompt in Terminal. The Code tab of the Claude desktop app doesn't run statuslines, so Claude Usage includes a small Claude Code plugin that draws the same bars above the prompt there. Nothing is installed until you press Install.

The Claude Usage statusline in a terminal: the model, the folder, context at 5%, a session bar at 0% resetting in 4 hours 31 minutes, and a weekly bar at 7% resetting Monday at 5:59 PM.
In Terminal: the bundled statusline, under Claude Code's prompt.
The Claude desktop app's Code tab, with session and weekly usage bars in a band just above the prompt: session 0% resetting in 4 hours 32 minutes, weekly 7% resetting Monday at 5:59 PM.
In the Claude desktop app: the same bars, above the prompt in the Code tab.

The fine print

  • Session and weekly limits, reset countdowns, and pace projections
  • claude.ai's per‑model weekly limits, and extra usage spend when it's on
  • Claude service status in the popover
  • Adjustable warning and critical thresholds
  • Model breakdown covers Claude Code; claude.ai doesn't report which model a chat used
  • Never reads or touches Claude Code's credentials
  • Never writes into ~/.claude itself — the statusline is yours to install, and the desktop plugin goes in through Claude Code's own command when you press Install
  • Native Swift app for Apple silicon, notarized by Apple
  • No built‑in updater yet — new versions are posted here
  • Free — no account with us, no analytics, no telemetry

Claude Usage is an independent app. It is not made by, affiliated with, or endorsed by Anthropic. Claude and Claude Code are Anthropic products.