Flow metrics and Monte Carlo delivery forecasts from a Jira CSV export, entirely in the browser. Upload an export (or load the bundled sample data) and get lead time scatterplots, throughput, cumulative flow, aging WIP, and probabilistic answers to "when will these items be done?" and "how many items by that date?" — with no plugin, no admin install, and no data leaving your machine. The equivalents are paid Jira marketplace apps or Python notebooks; this is the zero-install version a PM can use in 30 seconds.
Live demo: flowcast-a0v.pages.dev
npm install
npm run dev
npm test runs the Vitest suite; npm run gen:sample regenerates the bundled
sample dataset deterministically.
Flowcast bootstraps from your own history: it samples random days from your recent daily throughput (zero days included) 10,000 times, and reads the 50th to 95th percentiles off the simulated outcomes. No story points, no velocity — only items that actually finished on days that actually happened. The full method, its assumptions, and its failure modes are in METHODOLOGY.md.
- Lead time, not cycle time. CSV exports carry no status history, so time waiting in the backlog is included and time-in-progress is unknowable.
- Assumes system stability. The forecast projects the recent past forward; team changes, scope shifts, and trending WIP all break that assumption.
- Garbage in, garbage out. Stale tickets, bulk closes, and imported backlogs flow straight into the percentiles.
The app is a static build — any static host works. On Cloudflare Pages: build
command npm run build, output directory dist, no functions needed.
- Daniel Vacanti, Actionable Agile Metrics for Predictability — the book behind most of the conventions used here.
- Troy Magennis / Focused Objective — spreadsheet-based Monte Carlo forecasting tools.
- DeloitteDigitalUK/jira-agile-metrics — a Python CLI covering similar ground against the Jira API.
- lunivore/montecarluni — Monte Carlo forecasting from Jira exports, on the JVM.
MIT licensed. © Marta Davidoff.