Cost per business unit
The metric that drives AI pricing: not total spend, but cost per report, per candidate, per post.
7-day spend
$24.18
API calls
3902
Avg / report
$0.031
Anomalies
1
Spend over time
Cost per unit
Match report
412 calls
$0.031
/ report
Resume screen
286 calls
$0.018
/ candidate
Marketing gen
1240 calls
$0.0094
/ post
Brain extract
980 calls
$0.0011
/ insight
Recent API calls
| Time | Feature | Model | Tokens | Cost |
|---|---|---|---|---|
| 09:41:02 | Match report | claude-sonnet-4-6 | 5,960 | $0.031 |
| 09:40:55 | Marketing gen | claude-sonnet-4-6 | 1,592 | $0.0094 |
| 09:39:12 | Match reportInput 9x above average for this feature | claude-sonnet-4-6 | 40,300 | $0.124 |
| 09:38:01 | Resume screen | claude-sonnet-4-6 | 3,600 | $0.018 |
| 09:36:44 | Brain extract | claude-haiku-4-5 | 2,080 | $0.0011 |
Cost observability for AI products: an Electron menubar app intercepts API calls system-wide via a local HTTPS proxy into SQLite, and a Next.js dashboard rolls spend up by business unit.
Any app's API callLocal HTTPS proxySQLite captureRoll-upDashboard + anomalies