Cost Cockpit
The financial hub — AI spend and GitHub Actions consumption unified per repo, with drill-down to every run, commit, and PR that produced the numbers.
Table of Contents
💰 Cost Cockpit
One financial surface for the whole fleet: Claude Code spend (real CI cost from run logs + opt-in local estimates) and GitHub Actions consumption (minutes, waste, effectiveness), unified per repo. Every number drills down — to the /ai-usage/ audit ledgers, the /actions/ workflow table, and the runs, commits, and PRs at the source.
$13.52
Claude CI spend / 14d
$1132.02
Local est. / 14d
103.7h
Actions consumed / 14d
53.7h
Actions wasted (52%)
433
Claude runs · 547 turns
127+143
Claude commits + PRs
Cost by repo — the unified ledger
Actions minutes from the Actions API; Claude cost
from claude-code-action run logs. Row links: repo → GitHub,
runs → the repo's Actions tab, 🧮 → AI ledger, 📈 → workflow drill-down.
| Repo | Claude $ | Claude runs | Turns | Actions min | Waste min | Effective | Commits/PRs | Drill down |
|---|---|---|---|---|---|---|---|---|
zer0-mistakes |
$0.0 | 13 | 0 | 1837.6 | 1114.5 | 38.9% | 19/27 | runs · 🧮 · 📈 |
law-ai |
$0.0 | 5 | 0 | 1802.5 | 1500.8 | 16.7% | 11/13 | runs · 🧮 · 📈 |
lifehacker.dev |
$2.0075 | 51 | 55 | 788.9 | 90.5 | 87.7% | 67/71 | runs · 🧮 · 📈 |
it-journey |
$0.0 | 6 | 0 | 577.6 | 25.9 | 94.5% | 5/4 | runs · 🧮 · 📈 |
bashconsultants |
$5.5124 | 9 | 260 | 426.4 | 381.1 | 10.6% | 10/4 | runs · 🧮 · 📈 |
wtd |
$0.0 | 2 | 0 | 239.5 | 0.2 | 99.4% | 1/2 | runs · 🧮 · 📈 |
gitorio |
$3.1865 | 100 | 83 | 123.0 | 18.6 | 82.7% | 6/7 | runs · 🧮 · 📈 |
barodybroject |
$0.0 | 80 | 0 | 103.5 | 51.4 | 45.9% | 0/4 | runs · 🧮 · 📈 |
edgar-data-parse |
$0.0 | 2 | 0 | 50.5 | 1.5 | 96.8% | 0/0 | runs · 🧮 · 📈 |
ai-seed |
$0.0 | 72 | 0 | 47.1 | 4.1 | 87.8% | 0/0 | runs · 🧮 · 📈 |
bashcrawl |
$0.0 | 1 | 0 | 39.7 | 2.5 | 93.7% | 1/1 | runs · 🧮 · 📈 |
aieo |
$0.0 | 24 | 0 | 39.0 | 4.1 | 86.8% | 1/0 | runs · 🧮 · 📈 |
bamr87.github.io |
— | — | — | 24.0 | 0.0 | 100.0% | — | runs · 🧮 · 📈 |
githubai |
$2.8099 | 66 | 149 | 22.5 | 8.4 | 51.1% | 0/1 | runs · 🧮 · 📈 |
git-with-the-program |
$0.0 | 0 | 19.2 | 7.4 | 61.5% | 1/1 | runs · 🧮 · 📈 | |
cv-builder-pro |
— | — | — | 17.4 | 1.1 | 93.4% | — | runs · 🧮 · 📈 |
README |
$0.0 | 0 | 16.7 | 0.4 | 97.5% | 2/2 | runs · 🧮 · 📈 | |
zpl-viewer |
$0.0 | 1 | 0 | 9.3 | 7.1 | 22.8% | 0/0 | runs · 🧮 · 📈 |
vscode-front-matter |
— | — | — | 6.8 | 0.0 | 100.0% | — | runs · 🧮 · 📈 |
zer0-image-generator |
$0.0 | 0 | 6.2 | 0.0 | 100.0% | 1/1 | runs · 🧮 · 📈 | |
drsai |
— | — | — | 5.4 | 0.0 | 100.0% | — | runs · 🧮 · 📈 |
cv |
— | — | — | 5.1 | 0.0 | 100.0% | — | runs · 🧮 · 📈 |
vs-sonic-pi |
— | — | — | 3.0 | 0.0 | 100.0% | — | runs · 🧮 · 📈 |
csv-vscoode |
$0.0 | 1 | 0 | 2.7 | 0.0 | 93.2% | 0/0 | runs · 🧮 · 📈 |
SCHEMA |
$0.0 | 0 | 2.6 | 0.0 | 100.0% | 2/1 | runs · 🧮 · 📈 | |
lawmode |
— | — | — | 1.5 | 0.0 | 100.0% | — | runs · 🧮 · 📈 |
wargames |
— | — | — | 1.2 | 0.0 | 100.0% | — | runs · 🧮 · 📈 |
zer0-pages-remote |
— | — | — | 1.2 | 0.0 | 100.0% | — | runs · 🧮 · 📈 |
djangoerp |
— | — | — | 1.2 | 0.0 | 100.0% | — | runs · 🧮 · 📈 |
zer0-pages |
— | — | — | 0.9 | 0.0 | 100.0% | — | runs · 🧮 · 📈 |
scripts |
— | — | — | 0.7 | 0.0 | 100.0% | — | runs · 🧮 · 📈 |
books |
— | — | — | 0.5 | 0.0 | 100.0% | — | runs · 🧮 · 📈 |
irony-works |
— | — | — | 0.3 | 0.0 | 0.0% | — | runs · 🧮 · 📈 |
1987 |
— | — | — | 0.1 | 0.0 | 100.0% | — | runs · 🧮 · 📈 |
archify |
$0.0 | 0 | — | — | — | 0/4 | runs · 🧮 |
Where the money goes
Claude cost by workflow
| Workflow | Runs | $ |
|---|---|---|
githubai/Claude Maintenance | 2 | $2.3352 |
gitorio/Factory: Gitorio Factory 2 | 14 | $2.094 |
bashconsultants/Preacher (weekly doctrine) | 2 | $1.85 |
bashconsultants/Content gardener | 2 | $1.8484 |
bashconsultants/Content review (weekly) | 2 | $1.814 |
lifehacker.dev/Factory: Issue Factory 2 | 1 | $1.2407 |
lifehacker.dev/Factory: Issue Factory 1 | 2 | $0.7668 |
gitorio/Factory: Gitorio Factory 3 | 2 | $0.6422 |
githubai/Claude Review | 3 | $0.4747 |
gitorio/Factory: Gitorio Factory 1 | 2 | $0.4503 |
Actions minutes by type
| Type | Minutes | Effective |
|---|---|---|
| ci | 2420.5 | 37.4% |
| other | 1668.9 | 44.7% |
| ai | 762.6 | 92.8% |
| automation | 547.3 | 19.5% |
| deploy | 349.2 | 68.1% |
| security | 315.7 | 36.4% |
| dependencies | 102.7 | 92.4% |
| docs | 46.8 | 94.9% |
| release | 10.0 | 100.0% |
Daily burn
Claude $ / day
| Day | Runs | $ |
|---|---|---|
| 2026-09-02 | 30 | $0.1451 |
| 2026-09-03 | 43 | $1.4293 |
| 2026-09-04 | 37 | $0.1578 |
| 2026-09-05 | 31 | $1.2373 |
| 2026-09-06 | 33 | $0.1501 |
| 2026-09-07 | 76 | $4.8511 |
| 2026-09-08 | 8 | $0.4811 |
Actions min / day
| Day | Runs | Min |
|---|---|---|
| 2026-09-02 | 156 | 136.1 |
| 2026-09-03 | 317 | 264.5 |
| 2026-09-04 | 282 | 382.7 |
| 2026-09-05 | 344 | 371.9 |
| 2026-09-06 | 550 | 2246.9 |
| 2026-09-07 | 612 | 2087.0 |
| 2026-09-08 | 37 | 45.1 |
Sources & audit trail
Every figure traces to public infrastructure — the cockpit is a view, never a claim:
- Claude CI cost —
_data/ai_usage.yml, refreshed daily byfleet-pulse.yml; per-run cost scraped fromclaude-code-actionlogs, each ledger row on /ai-usage/ links to its run. - Actions consumption —
_data/actions_usage.yml, refreshed daily byfleet-pulse.ymlvia the Actions API; per-workflow drill-down on /actions/. - Optimization loop — the worst offenders are merged with the failing-workflow signal into one ranked queue (
dash-gen remediate) and fixed by the Opusdoctorjob offleet-pulse.yml, which opens a draft PR in the owning repo or falls back to a hub issue. - Local sessions — token-true estimates from each machine’s ledger, published only by a deliberate local
tools/dash-gen ai-usagerun (/ai-activity/ for the local-only view).
Claude data 2026-09-08 06:34 UTC · Actions data 2026-09-08 06:31 UTC