tokenfficiency

Compare your AI token spend to the code you actually ship.

Get started

Run this in your terminal. Auto-detects your coding agents and publishes a shareable report.

npx tokenfficiency

Optional: gh auth login for GitHub shipping stats. --no-post for local only.

What it measures

Code shipped

  • ·Merged PRs + direct pushes via GitHub API
  • ·Per-file noise filtering (lock files, generated code)
  • ·Merge commits and bot activity excluded

Token usage

  • ·Claude Code via ccusage
  • ·OpenCode via @ccusage/opencode
  • ·Codex via @ccusage/codex
  • ·Pi via @ccusage/pi
  • ·Amp via @ccusage/amp
  • ·Per-agent and per-repo breakdown

Grading

Based on cost per line of shipped code (after noise filtering):

S
< $0.001/line
A
< $0.005/line
B
< $0.01/line
C
< $0.05/line
D
< $0.10/line
F
> $0.10/line

Recent reports

@gracelliuC
+40,286 lines · $711.77
@mclenhardA
+78,152 lines · $207.75
@wilw-vercelC
+7,133 lines · $178.54
@yashk2005C
+196,091 lines · $8757.32
@simhskalA
+136,068 lines · $137.80
View all 11 reports →