Bill-y.
AI coding spend, measured by workflow, not just volume.
The idea
See whether AI is used well, not just heavily.
Engineering teams now spend hundreds of dollars per developer each month on AI, and the bills swing without warning. Counting tokens tells you how much. It does not tell you whether they were spent well. Bill-y splits that spend across the phases of building software, so a leader can finally see the shape of how their team works with AI, and where the hidden waste is.
What it does
A monitoring layer for AI efficiency.
Phase attribution
Splits token spend across understand, design, build, test, and review, so cost maps to the real work.
Efficiency, not volume
Measures cost per outcome instead of raw usage, the one metric that survives being gamed.
Built for Claude Code
Instruments the richest telemetry surface there is, per step, per tool, per subagent.
Private by design
Runs locally. An efficiency coach, not a surveillance dashboard.
Catches hidden waste
The distribution is a leading indicator of the rework that vibe-coding pushes downstream.
FinOps for AI
Arrives as token bills become a naked, separately metered line on the invoice.
Editions
Two products, one idea.
Bill-y for Enterprise
For engineering teams and the leaders who own their AI spend.
Request early access →Bill-y for Developers
A completely free edition for individual developers. Coming soon.
How it works
Observe, measure, attribute.
Capture real usage
Read live usage from your AI coding tools, per step and per tool, with no change to how the team works.
Classify each turn
Sort every turn into a workflow phase and tally the tokens and time it actually consumed.
See the shape
Roll it up into a clear picture of spend by phase, by project, and over time.
Get Bill-y
Request early access.
Bill-y is in active development. Self-serve pricing and checkout are coming. For now, tell us about your team and we will get you set up.