Free tool · Cratesignal

Plan a reorder without inventing a PO size.

Drop in current stock, sell-through rate, and lead time. The tool runs deterministic math server-side — days of cover, reorder point, safety stock, recommended reorder date, and recommended reorder quantity — and feeds the same numbers to the AI narrative as invariants. The model can't invent a different safety stock figure or a different PO unit count.

No signup
Stateless
≥60 days-of-supply default
Reorder brief
Six required inputs. The deterministic math runs server-side; the AI narrative is generated from the citation numbers, not invented from scratch.

% of current stock you expect to move per window. 0–100.

Window the rate applies to — converted to per-day internally.

Whole units on hand — drives days of cover.

Supplier replenishment time, in whole days.

Sets the cohort baseline band (low / typical / high) shown on the result panel.

What one PO should cover — defaults to 60 (30-day lead + 30-day buffer).

0.30

Extra days on top of lead time: 30% — adds ~14 cover days.

Labels the baseline — kitchen, pet, both, or other.

Picks the sell-through cohort bucket (low / mid / high / premium).

Stateless: nothing saved.

Result

Your reorder plan will appear here.

Drop in current stock, sell-through rate, and lead time — the tool computes days of cover, reorder point, safety stock, recommended reorder date and quantity, and a one-paragraph AI narrative grounded in the citation numbers from Cratesignal's 2026-Q3 cohort baseline.

How the data is wired

The recommended reorder date and quantity are computed deterministically at the route boundary — never requested from the model. The math is dailySales = currentStock × sellThroughPct / 100 / windowDays for the run-rate, daysOfCover = currentStock / dailySales for cover, and reorderPoint = ceil(dailySales × leadTime × (1 + safetyMultiplier)) for the trigger. The recommended PO quantity covers the gap from current cover to the user-set target days of supply, clamped to zero when you're already over-covered.

Sell-through baselines come from Cratesignal's 2026-Q3 kitchen + pet cohort, indexed by price bucket (sub-$15: 12% / month, $15–30: 9%, $30–50: 7%, $50+: 4%). Lead-time bands come from the same snapshot (Asia sourcing 30 / 45 / 65 days, EU 10 / 14 / 22, US 3 / 6 / 12). Each number is rendered back on the result panel as a citation chip, and the AI narrative is given the same numbers as invariants — it does not get to invent or recompute them. The plain-text disclosure on the result panel says exactly where the figures come from.

Voice rules are folded into the system prompt at src/lib/inventory/prompt.ts : no emojis, no superlatives, no aspirational copy, every sentence cites at least one numeric. If you tweak the model — temperature, vendor, baseline year — update the snapshot date on the result-panel footer so the citation provenance stays honest.

For questions or to flag a baseline drift, write to cratesignal-3@polsia.app.