OpenBoxes
Inventory, warehouse and healthcare supply-chain workflows. A deployment can export product/site stock into this page’s CSV schema.
EPL-1.0 licenseRADIOGX / BYOK AI WORKBENCH
Healthcare supply chain intelligence
See where stock, shelf life and delivery assumptions collide. Compare a baseline with disruptions, then ask your AI to explain the risks and synthesize a review plan.
Loading the planning workbench…
01 / INVENTORY
Your planning inputs
Use one consolidated product/site row. Enter quantities in a consistent base unit for that product. “Near-expiry units” belong to stock that is not held; the rest is assumed to last beyond the horizon. Scroll the table horizontally to edit all fields.
| Product / base unit | Site | On hand | Held | Demand / day | Lead days | Safety days | Incoming | Near-expiry units | Expiry in days | Cold chain | Cost / unit | Edit |
|---|
Input format and local storage
Import a CSV with the template headers, a JSON inventory array, or an exported RadioGx supply input/report document. Import is capped at 1 MB and 100 rows. Each quantity is a whole base unit except daily demand and unit cost. “Expiry in days = 0” means already expired. Cost uses one currency of your choice across the workbook.
Imported data is unverified. The page does not keep inputs after refresh unless you choose “Save inputs on this device”. Saving stores inventory and scenario assumptions, never your AI key. AI settings control key storage separately.
03 / TRANSPARENT CALCULATIONS
Supply risk dashboard
- SupplierOne entered shipment
- Depot / storeUsable stock after holds
- Quality reviewExpiry & scenario hold
- Site demandDaily stock simulation
- Human reviewOrder gap & AI synthesis
All results below are derived from your inputs. The model does not place orders. Day 0 is today; stock expires at the start of the stated expiry day. Quantities across different products are not added.
| Product / site | Usable now | Static days cover | Reorder point | Planning order gap | Expiry loss | Unmet demand | First short day | Flags |
|---|
End-of-day stock. Baseline removes the demand, delay and cold-chain shocks. A receipt is included only when it arrives within the horizon; suggested orders are excluded.
Model formulas, limits and assumptions
Static days cover = usable stock now ÷ adjusted daily demand. It excludes future receipts and future expiries; the simulation includes both. A zero-demand assumption displays “No demand”. Planning order gap is a heuristic stock policy; it is not an optimized purchase or a clinical allocation.
04 / BYOK AI
Inference & synthesis
Your selected AI reviews the computed baseline and scenario. Ask it to separate observations, causal hypotheses, operational options and missing evidence. It has no access to a live supply system or external evidence through this panel.
Use public or de-identified information. The report contains product names, sites, quantities, costs and assumptions; these are sent to your chosen AI provider when you generate a response.
No AI report generated.
Calculate your inventory scenario, connect your AI key, then generate a research planning narrative.
05 / INFRASTRUCTURE
Open-source architecture references
This browser workbench implements its own transparent planning rules and uses the shared RadioGx BYOK provider layer. The projects below are verified integration references for a future operational deployment. Their servers and optimization engines are not bundled or executed here.
OpenLMIS
Reference distribution for last-mile health commodity information systems. An integration should map stock, consumption and fulfillment records through a trusted server.
AGPL-3.0 licenseGoogle OR-Tools
Future constrained routing, scheduling and optimization services. This page’s order-gap calculation is a heuristic; no OR-Tools solver is loaded.
Apache-2.0 licenseDeployment handoff and data contract
The exported input JSON is the integration boundary: schema radiogx-supply-v1, items array and scenario object. A trusted connector should reconcile units, product/site identifiers, reservations, timestamps, cost currency, lot-level expiry and receipt confidence before import. Operational credentials stay in your server integration; this page does not request them.
Repository roles and linked licenses checked 19 October 2026. No source from these three repositories is copied into this module. See the supplied source documentation for the complete contract and verification steps.
About this tool
What is Supply chain intelligence?
Runs in your browserBYOK AI step
Healthcare supply chain intelligence is a browser planning workbench that projects stock with first-expiry-first-out consumption, delivery delays, demand surges and cold-chain holds, compares baseline and disruption, and asks your own AI to synthesize the planning risks.
Key features
- Editable inventory with clearly synthetic examples
- CSV and JSON import and export with validation
- Baseline versus demand surge, delivery delay and cold-chain hold scenarios
- Daily FEFO expiry, days cover, safety stock, reorder point and order gap
- Stock projection chart and per-product risk list
- BYOK risk synthesis; no procurement action is ever taken
How it works
- Load the synthetic examples or import de-identified inventory
- Set the horizon and disruption assumptions
- Calculate baseline and scenario projections
- Ask your AI to interpret the results and export the report
Frequently asked questions
Does the supply chain tool place orders?
No. It calculates planning signals such as days cover and order gaps for research and teaching. No procurement or supplier connection is made.
What is FEFO?
First expiry, first out: stock with the earliest expiry is used first. RadioGx simulates it daily so expiring units are removed from usable stock.