Mock Recall & Trace Simulator: Prove a 4-Hour Recall, On Demand
Pick a suspect lot, trace it forward to every customer and shipment and backward to its inputs, time the whole thing, and produce the documented recall package auditors require — with a manager sign-off before any notice goes out.
A web tool where you import your genealogy and shipment data, pick a suspect lot, watch it trace forward to every finished lot, customer, and shipment it reached (and backward to its inputs), see a timed trace and a made-vs-shipped-vs-on-hand quantity reconciliation, get a manager to approve the recall scope and notification list, then export a complete recall package plus customer-notice drafts via Resend.
Before you start
- A Supabase account (free)
- A Vercel account (free)
- A Resend account (free)
- A CSV of your lot genealogy / consumption data and a CSV of your shipment records (lot → customer → ship date/qty)
- Claude Code or any AI coding agent
The problem this kills
When the auditor asks "show me you can trace a suspect lot to every affected customer in four hours," the honest answer in most plants is a held breath and a stack of spreadsheets. Genealogy lives in one export, consumption in another, shipments in a third — and stitching them together by hand under pressure is exactly when mistakes happen. Miss one finished lot that consumed the suspect input, and you miss the customers it reached. Over-scope it, and you recall product that was never at risk and burn trust you can't get back.
The certifications you live under — SQF, BRC, FDA — don't just want a recall plan; they want proof you can execute one, fast, with the timing and documentation to show it. A mock recall is how you prove it. But running one by hand is so painful that most teams do it once a year, dread it, and never quite trust the result. You do not need to be a developer to build a tool that runs the whole trace in seconds, times it, reconciles the quantities, and hands you the package — every time.
What you'll build
A simple internal web tool for your quality and food-safety team. You import two files: your lot genealogy / consumption data (which input lots went into which finished lots) and your shipment records (which finished lot went to which customer, when, and how much). You pick a suspect lot. The tool traces forward through every finished lot that consumed it — and every lot those fed — all the way to the customers and shipments it reached, and backward to the inputs and suppliers it came from. It time-stamps the trace so you can measure your recall window against the four-hour target, and it reconciles quantities: how much was made, how much shipped, how much should still be on hand. Your quality manager reviews the traced scope and the customer-notification list and clicks Approve — and only then does the tool export a complete recall package (the trace tree, the affected-customer list, the timing record, the quantity reconciliation) and draft customer-notice emails via Resend. Nothing leaves the building until a person signs off.
What's inside the Implementation Plan
The downloadable plan is a step-by-step file you paste into an AI coding agent. It opens by interviewing you about your business — how lots and batches are numbered in your plant, how your genealogy and shipment exports are actually shaped, how deep your bill-of-materials trees go, your typical and peak volumes, and the messy edge cases like rework, repacks, blended lots, and split shipments. It reads a short spec back to you for a thumbs-up, then builds the tool around your lot numbering and your data — not a generic template. From there it walks the agent through the data model, the two CSV imports, the forward-and-backward trace engine, the timing and quantity-reconciliation logic, the manager approval gate, and the recall-package and customer-notice exports. Every step ends with a ready-to-copy prompt.
The governance it includes (this is the point)
This isn't a toy. The plan builds in the controls a real quality function needs: login so only your team can use it, row-level security so people only ever see their own organization's lots and customers, a complete audit trail of every trace run, scope change, and approval (who, what, when), a hard human-approval gate so no recall scope is finalized and no customer notice is drafted until your quality manager approves, and duplicate guards so the same genealogy or shipment file can't be imported twice and skew the trace. The whole tool exists to make a careful, defensible decision easy and fast — the AI assembles the trace, a person owns the call.
Who it's for
Quality managers, food-safety and compliance leads, and anyone who owns mock recalls and trace exercises under SQF, BRC, or FDA. If you can describe how your lots are numbered and where your genealogy and shipment data live, you can build this.
You've got this — open the plan, paste the first prompt, and you'll be running your first mock recall this afternoon.