%20--%3e%3crect%20width='1024'%20height='1024'%20rx='224'%20fill='%2316181d'/%3e%3cpath%20d='M%20292.8%20731.2%20A%20310%20310%200%201%201%20731.2%20731.2'%20fill='none'%20stroke='%238a909b'%20stroke-width='92'%20stroke-linecap='round'/%3e%3cpolygon%20points='700.1,329.6%20539.8,540.7%20484.2,483.3'%20fill='%23388bfd'/%3e%3ccircle%20cx='512'%20cy='512'%20r='44'%20fill='%23f2f3f5'/%3e%3c/svg%3e){kind=small}
Token Gauge
Claude Code usage in your menu bar — not another browser tab
If you run Claude Code every day, you know the interruption: mid-refactor, you wonder how much of your rolling window is left, open a dashboard, lose your flow, and come back hoping you weren't already at 94%. Anthropic enforces limits that rarely line up with how you work — a 5-hour rolling window, an overall weekly cap, and per-model weekly limits on top — and none of them shows up next to your editor unless you go looking.
Every clock, one glance
Token Gauge is a small, native menu-bar app for Mac — and the same app, on the same license, runs in the Windows system tray. It reads Claude Code's usage on your machine and shows live 5-hour and weekly percentages, color-coded green, amber or red, plus a countdown to the next reset in your local timezone. Since v1.1 each model's own weekly limit gets its own bar too — a separate "Fable — Weekly" gauge, for example, because a model's cap can run dry days before your overall week does. Pin Claude as your primary agent and the menu bar shows the number that matters without opening the popover at all.
What you get for Claude Code
The free tier includes full Claude Code tracking — no license key, no Token Gauge account required to start. You get the live 5-hour window, the weekly cap and per-model weekly limit bars with reset countdowns, per-model token and cost tables, and a usage-credits spend row when your account has credits enabled — all computed on your machine. Threshold notifications let you set a nudge before you burn the window during a long session, including on per-model limits. Heads up: this works on Claude Pro, Max, Teams, Enterprise and the Console — the free claude.ai plan has no usage endpoint, so it reads blank.
Local reads, honest scope
We're explicit about what Token Gauge does and doesn't do. It reads usage from Anthropic's own authenticated usage endpoint using credentials already on your system — never your repositories, prompts or conversation logs. That read is authoritative and doesn't consume any of your quota. The only network calls Token Gauge makes are each provider's usage endpoint plus one license check; it isn't fully offline, but your code and prompts never leave the machine because there's nowhere for that data to go.
Privacy in plain terms: Token Gauge asks Anthropic "how much have I used?" with credentials you already have. No analytics, no tracking pixels, no telemetry. It gives you visibility and timing — it does not raise your limits or change what you pay.
Claude Code, straight answers.
Anthropic meters Claude Code on a rolling 5-hour window: every message counts against it, and it resets 5 hours after your first message in that window, not at a fixed clock time. Token Gauge reads that window directly from Anthropic's usage endpoint and shows it as a live percentage, color-coded green, amber or red, with a countdown to when it resets in your local timezone.
Alongside the 5-hour window, Anthropic enforces an overall weekly cap, and since Token Gauge v1.1 each model gets its own weekly limit too — shown as a separate bar, like "Fable — Weekly". A model's own limit can run out days before your overall week does, so Token Gauge tracks both the shared weekly cap and each model's limit with their own reset countdowns.
No. Token Gauge is a meter, not a booster — it shows your usage and reset timing so a limit never surprises you mid-task. It does not raise your 5-hour window, add weekly quota, unlock higher per-model limits, or change what you pay Anthropic. Anthropic sets and controls every limit; Token Gauge only reads and displays it.
Yes. Full Claude Code tracking — the 5-hour window, weekly cap, per-model weekly limits, per-model token and cost tables, and threshold alerts — is free, with no license key and no Token Gauge account required. The $7.99 lifetime license only unlocks tracking for the other agents: Codex, Cursor, Copilot and Antigravity.
Not for usage data. Token Gauge reads Claude Code's usage from Anthropic's authenticated usage endpoint, which is only available on Claude Pro, Max, Teams, Enterprise and Console plans. The free claude.ai plan has no usage endpoint, so those fields read blank — this is an Anthropic limitation, not something Token Gauge can work around.