AI Agent Usage
A free native macOS app that shows exactly what your AI coding agents cost you — Claude Code, Cursor, OpenAI Codex, and Gemini CLI, all in one place, fully on-device, no account required. It started as cost tracking, but the real payoff is parsing each run's token details to see where the spend actually goes.
View on the Mac App Store →I use a handful of AI coding agents every day, and at some point I realized I had no real idea what any of them were costing me. The bills showed up, the numbers went up, and I just kind of nodded along. So I built AI Agent Usage — a native macOS app that puts Claude Code, Cursor, OpenAI Codex, and Gemini CLI in one place and tells me exactly what each one costs. It's free on the Mac App Store, runs entirely on-device, and there's no account to create.
Why I built it
The honest answer is that I wanted to stop guessing. When you've got four different agents running against four different pricing models, the spend gets murky fast. I wanted one screen that says, plainly, here's what you spent and here's where it went — without shipping my usage data off to someone else's server. Everything stays on your machine. No login, no telemetry, no signup flow standing between you and the number you actually came for.
How it works
It started as straightforward cost management: calculate the spend per run, add it up, done. That part is genuinely useful, but it turned out to be the boring half. The interesting half is the insights.
The app parses the raw token details of each individual run, so instead of one lump-sum number you can see the shape of the cost:
which runs are quietly expensive and which are basically free
where the tokens are piling up — context, output, the long-running sessions
what's actually worth optimizing versus what just feels like it should be
That last point is the one I keep coming back to. It's easy to assume you know where your spend is going. Looking at the real token breakdown is usually a little humbling, and occasionally it changes how I work.
Where it's headed
For now I'm keeping it tight: more agents, sharper breakdowns, and making the insights even easier to act on at a glance. It's the tool I wanted for myself, and it stays on-device by design — your usage is your business. If you're running coding agents and you've never really looked at the bill, give it a try and see what surprises you.