Track AI coding agent usage from your MacBook notch
See how much of your Claude, Codex, Kimi, GLM, and DeepSeek subscription you've used — and how long until each window resets — without opening a single dashboard. Vibe Island reads your existing CLI sessions and surfaces it all in one notch panel.
One panel, five providers
Most developers juggle two or three AI coding subscriptions. Vibe Island brings every provider's quota into a single notch view: Claude, Codex, Kimi Code, GLM, and DeepSeek. Each provider gets its own quota bar with current usage, remaining window time, and reset countdown.
Real-time, not cached
Quota updates as you use the CLI. Send a Claude Code message and your notch bar moves immediately — no manual refresh, no waiting for a polling cycle. The same loop that powers session tracking drives the quota display, so it's always current.
Window types you actually have
Different providers slice their limits differently — 5-hour rolling windows, 7-day cumulative caps, monthly token budgets. Vibe Island shows whichever window matches your plan, with the exact reset time so you can plan your work around it.
Zero config detection
The first time you launch Vibe Island, it detects every supported CLI you have installed and starts tracking usage automatically. No keys to copy, no provider URLs to configure, no enterprise IT tickets. Just install and look.
Click to see the breakdown
Hover over the notch and you'll see each provider's full usage card — current consumption, this window's reset, and provider plan tier. Click to jump to the official billing page for that provider when you want the full ledger.
Check your Codex and Claude Code usage at a glance
The usual way to check Codex or Claude Code usage is to run a command, read a wall of output, or open a provider dashboard in the browser — every time you wonder how much you have left. Vibe Island removes that loop entirely. Your current usage, the percentage of your limit consumed, and the time until the window resets sit in the notch, always visible. When you're deep in a session and want to know whether you can start one more big task, the answer is already at the top of your screen — no context switch, no waiting on a refresh.
Token and cost breakdown, by day and model
Remaining quota tells you when you'll be cut off; it doesn't tell you where your usage went. Vibe Island brings the full breakdown into the same panel — input, output, and cache tokens, estimated dollar spend, and a per-model split — across the current day and the month so far. You can see that a single model is eating most of your budget, or that cache reads are quietly carrying the load, and adjust how you work before the bill (or the rate limit) makes the decision for you.
Common questions
Which AI coding agents does Vibe Island track usage for?
Vibe Island tracks usage and quota for Claude, Codex, Kimi, GLM, DeepSeek subscriptions. Each provider's remaining quota, window reset time, and limit status appears in the notch panel.
Do I need to enter API keys to see my usage?
No. Vibe Island reads your existing CLI login session directly — the same authentication you use to run Claude Code, Codex, or any other supported agent. No extra credentials, no API keys, no additional accounts.
How accurate is the usage data?
Vibe Island reads from each provider's authoritative source — the same data the CLI itself uses. There is no estimation or guessing. When you see 65% used in the notch, that's exactly what the provider reports.
Will Vibe Island warn me before I hit my limit?
Yes. The notch panel shows a real-time quota bar that turns warning colors as you approach the limit. You can also see when the window resets, so you know whether to wait or switch to a different provider.
Does usage tracking work for team or enterprise plans?
Yes, as long as the underlying CLI supports the plan. Vibe Island reads from whatever account you're logged in as — personal, team, or enterprise — and shows the corresponding quota.
How do I check my Codex or Claude Code usage?
Glance at your Mac notch. Vibe Island shows your current Codex and Claude Code usage, the percentage of your limit you've used, and the time until your window resets — no command to run, no dashboard to open. It's the fastest way to check usage without breaking your flow.
How does the 5-hour rate-limit window work?
Claude Code and Codex meter usage on a rolling 5-hour window, on top of weekly and monthly caps. Vibe Island shows exactly which window applies to your plan and counts down to the reset, so you can see whether to keep going, wait it out, or switch to another provider.
Does it track token cost, not just limits?
Yes. Beyond remaining quota, Vibe Island brings token and cost breakdowns into the notch — input, output, and cache tokens, estimated dollar spend, and a per-model split — across the day and month, so you can see where your usage actually goes.
Pair it with the rest of Vibe Island
Usage tracking works alongside AI coding session monitoring, terminal jump, and permission approval — every Vibe Island feature in one notch. Try it on Claude Code, Codex, Kimi Code, or DeepSeek. Works on remote servers via SSH too.
Vibe Island Track every AI coding agent's usage from your notch
macOS 14+ · Intel & Apple Silicon
Download Free Trial2-day free trial · all features included · no card required