jank.ai
ROI calculator — Jank cloud reports
Each /jank_cloud run replaces hours of senior-QA contractor work and compresses calendar wait from days to minutes. Configure your scope and cadence — see exactly what you save.
Configure your run
cloud/server/value.js. Plugin calculator →Single run
Cumulative — your team
| Window | Runs | Hours | Value | Wait days saved |
|---|---|---|---|---|
| Per run | 1 | — | — | — |
| Per month | 22 | — | — | — |
| Per quarter | 66 | — | — | — |
| Annualized | — | — | — | — |
Why these numbers are defensible
Single formula, single page — open cloud/server/value.js. ~150 lines. Every constant has a default and an env override. Auditable in 5 min.
$50/hr is BELOW market. Glassdoor 2025 US median for senior-QA contractors is $75-120/hr. We deliberately undercut so the cost figure cannot be called inflated. Override JANK_HOURLY_USD at deploy if you want a different baseline.
Per-line constants are conservative. 10 min/finding (industry mean for a finding with title + repro + fix + routing); 30 min/persona session (real persona feedback takes 20-40 min); 20 min/flow (Cypress/Playwright record + replay + verify); 30 min/subpage (full audit minus persona pass); 0.5 min/link (scripted-rate, not by-hand). Each is below contractor-survey medians.
Calendar wait is single-tester math by default. Most teams don't have a QA crew on standby — the realistic wait is "how long until your next-available tester finishes this." Bump tester-pool size if your team has dedicated QA capacity ready to grab any incoming work.
Cumulative tally is real, not projected. Every cloud run writes to Firestore (jankAccounts/<email>.usage) atomically with idempotent retries. Pull your account's actual lifetime total at reports.jank.ai/admin#stats.
Full source-of-truth doc: ROI_CALCULATOR.md (12 sections + FAQ defending the math under every common challenge).