DEV Community

Jamie
Jamie

Posted on

I got tired of hitting AI rate limits mid-task, so I built a macOS menu bar monitor for it

If you use more than one AI tool daily — Claude, Codex, Cursor, Gemini, Copilot — you've probably hit this:

You're in the middle of a deep coding session. Flow state. Everything clicking. Then: "You've reached your usage limit."

No warning. No countdown. Just a wall.

The worst part isn't the limit itself. It's that you had no idea it was coming. You could have paced yourself, switched providers, or front-loaded the heavy work. But the information wasn't there.

The Problem

Every AI provider handles limits differently:

  • Claude has session windows and weekly caps
  • Codex has credits that deplete at variable rates
  • Cursor has monthly request limits with fast/slow tiers
  • Gemini has RPM and daily limits
  • Copilot has monthly completions caps

None of them show you a unified view. You're expected to check each dashboard separately, remember reset times, and somehow mentally track where you stand across all of them.

For devs who use 3-4 of these tools daily, it's a real workflow problem.

What I Built

TokenBar is a macOS menu bar app that tracks your AI usage across providers in real time.

It sits in your menu bar and shows:

  • Session/weekly/credit meters with reset countdowns
  • Pace intelligence — are you on track to last until reset, or burning through too fast?
  • Budget mode — set a target number of days and track live
  • Threshold alerts — get notified at custom percentages before you hit limits
  • Status monitoring — see if a provider is having incidents

Supported Providers

Codex, Claude, Cursor, Gemini, Copilot, OpenRouter, Vertex AI, Augment, Amp, JetBrains AI, Ollama, Warp, Kimi, Kiro, and more (20+ total).

How It Works

TokenBar uses a mix of auth methods depending on the provider: CLI tokens, OAuth, API keys, browser cookies, and local file probes. Everything is processed locally on your Mac — no telemetry backend, no cloud sync.

Key technical details:

  • Native Swift/SwiftUI (not Electron)
  • WidgetKit widgets for desktop
  • Bundled tokenbar CLI for terminal automation and CI
  • JSON output for scripting
  • Refresh cadence from 1min to 30min (or manual)
  • Global keyboard shortcut

The Feature That Changed My Workflow

The pace indicator is what made this worth building. It's not just "you've used 60%" — it tells you whether your current usage rate will get you to the reset window or not.

Seeing "deficit" at 2pm when your session resets at midnight means: slow down now, or switch to a different provider for the heavy stuff. That signal alone has prevented probably 30+ surprise limit hits for me.

Pricing

$4.99 one-time payment. No subscription. No recurring charges. You buy it once and it's yours.

🔗 tokenbar.site


If you've been frustrated by the same quota-surprise problem, I'd love to hear how you currently handle it. And if you try TokenBar, feedback is genuinely welcome — especially on which providers matter most to you.

Top comments (0)