If you build with AI tools daily, you've probably had this moment: you're deep in a session, you ask Claude one more thing, and you get hit with "you've reached your usage limit." No warning. No idea how close you were five minutes ago.
I kept running into this — across Claude, Cursor, Codex, Copilot, Gemini, OpenRouter — and I got tired of it. Each provider has its own dashboard, none of them update in real time, and switching between them mid-flow kills your concentration.
So I built TokenBar.
What it is
A native macOS menu bar app that shows your live AI usage across 20+ providers in one place. It sits in your menu bar, updates in real time, and tells you exactly where you stand before you hit the wall.
What it tracks
- Claude (Pro, Max, API)
- Cursor (all plans, including the new "unlimited but rate limited" Pro)
- GitHub Copilot
- OpenAI Codex
- Google Gemini
- OpenRouter
- Augment, Amp, JetBrains AI, Warp, Kiro, and a dozen more
For each one: current session usage, weekly/monthly limits, reset countdown, and a pace indicator that warns you if you're burning faster than usual.
Why menu bar
Because dashboards lose. The whole point is glanceability — one look up at the top of your screen, you know whether to keep going or wrap up. No tabs, no logins, no context switch.
Pricing
- $5 lifetime — basic (Claude + Cursor + Codex + Copilot)
- $10 lifetime — pro (all 20+ providers, pace indicators, incident detection, multi-account)
One-time purchase. No subscription. I'm a solo dev and I hate subscriptions as much as you do.
Try it
If you've ever lost a build session to a surprise rate limit, this is for you. Would love feedback from anyone who tracks usage across multiple AI providers — what else should I add?
Top comments (0)