Everyone who builds with AI eventually hits the same moment.
You're deep in a coding session. Claude is flying. You're feeling productive. Then you open your API dashboard and the number hits you like a bucket of cold water.
That happened to me. I don't want to talk about the exact number, but it was enough to make me stop and actually think about what I was doing.
The problem wasn't that I was spending money. The problem was that I had no idea I was spending it.
The dashboard problem
My first instinct was what everyone does: open the Anthropic dashboard. Check the usage graphs. Try to correlate the spikes with what I was working on.
But here's the thing about dashboards — they're designed for after-the-fact analysis, not real-time awareness. You go to a dashboard when something's already wrong. It's reactive, not preventive.
I tried setting up billing alerts too. Which helped, but it still didn't give me the in-the-moment feedback I actually wanted. I wanted to feel the cost as I was working, not find out about it 20 minutes later via email.
Why a menu bar?
I spend most of my day in my code editor. Terminal. Browser. The menu bar is the one UI element that's always visible regardless of what app I'm in.
When I'm writing a prompt, I want to glance up and see the token count growing in real time. When I'm about to send a massive context window to the model, I want to know before I hit send — not after.
A dashboard requires me to switch context. A menu bar widget doesn't.
It's the same reason CPU/RAM monitors have lived in the menu bar forever. You don't open Activity Monitor every five minutes to check if your computer is struggling. You just glance at the little graph and keep working.
Token usage is the same class of information — it's ambient, always-relevant, and you want it peripheral, not in your face.
What I built
So I built TokenBar — a macOS menu bar app that shows your LLM token usage in real time.
It hooks into your API activity and shows you a live counter right in the menu bar. Claude, GPT-4, whatever you're using. You can see your current session burn rate, your daily total, and you get a heads-up when you're approaching a threshold you set.
Nothing revolutionary. Just the information I wanted, in the place I was already looking.
The thing I didn't expect
Building it changed how I actually use AI.
When you can see tokens ticking up in real time, you start making different decisions. Do I really need to paste this entire file as context? Can I be more specific in this prompt? Is this follow-up question worth the cost or should I just test it myself?
It made me more intentional. Not stingy — I still use AI constantly. But intentional.
There's something about ambient feedback that changes behavior in a way that weekly reports never do. Same reason people spend less when they pay with cash vs. card — the friction of seeing it happen in real time is different from reviewing a statement.
For other solo devs
If you're building with AI on your own and you're not tracking your usage in real time, you're flying blind. The costs are real and they sneak up on you in the exact moments you're most focused.
You don't need a complex setup. You just need to see the number.
If you want to try what I built: tokenbar.site. It's a $5 one-time purchase, which is less than what one expensive mistake used to cost me in a day.
Top comments (0)