Demonstration site All companies, lots and records shown are fictional. BatchBeacon does not provide regulatory certification.
BatchBeacon

Food traceability · Batch investigation

Trace every lot.
Show every gap.

BatchBeacon links supplier lots, production runs, quality records, packaging and customer deliveries — so when a supplier calls, your team can scope the problem from recorded evidence in minutes, not days.

Try the live trace Request a demo
  • Scope comes from recorded relationships — never from a language model
  • Missing records and uncertainty shown explicitly
  • Recall decisions stay with your authorised people
SUPPLIER LOTPREMIXFINISHEDPACKEDDELIVERED PAP-117 PMX-0412 PMX-0415 PMX-0419 FB-0919 FB-0921 FB-0923 FB-0924 FB-1001 FG-0919 FG-0921 FG-0923 FG-0924 FG-1001 DLV-10231 DLV-10232 DLV-10240 DLV-10244 DLV-10249 DLV-10252 DLV-10270 100 cases unaccounted
VerifiedUnverifiedPossible Illustrative — fictional data
Recorded, not inferredEvery link cites the consumption, packing or shipment record behind it.
Gaps surfacedMass-balance shortfalls, unverified entries and missing certificates are listed, not hidden.
People decideOnly an authorised reviewer can approve recall scope — with a written justification.
Full audit trailEvery trace, question, extraction and decision is logged.

Product overview

One investigation workspace from supplier lot to customer dock

BatchBeacon sits on top of the records you already keep — receiving logs, MES consumption, packaging runs, QA checks and shipping — and turns them into an evidence graph your team can question.

Traceability Assistant

Authorised users ask plain-language questions — "which customers received product made with lot 2409-117?" — and get answers built from trace-engine results, with every record cited.

Batch Relationship Intelligence

Explains how a supplier lot connects to a premix, a finished batch, a packed lot and a customer — step by step, with the strength of each link.

Recall Scope Preparation

Lists potentially affected finished batches, packed lots, deliveries and stock on hand — tiered by evidence strength — ready for an authorised reviewer.

Quality Record Analysis

Extracts fields from certificates of analysis, inspection sheets and supplier letters, with a confidence score on each field and a human verification step.

Incident Report Generation

Structured investigation summaries: affected-lot lists, missing evidence, uncertainty statement and proposed next steps — clearly marked as a draft until reviewed.

Included modules

  • Ingredient lot register
  • Supplier records
  • Production batch tracking
  • Quality inspections
  • Customer deliveries
  • Relationship explorer
  • Incident workspace
  • Recall preparation reports
  • Audit logs

How the AI is constrained

The language model never decides what is affected

Scope is computed by a deterministic trace engine over your recorded relationships. AI is used to read documents and to word summaries — and its output is checked against the records before anyone sees it.

  1. 1Recorded relationshipsPostgreSQL: receiving, consumption, packing, shipments, stock counts
  2. 2Trace engineGraph traversal. Each link tagged verified, unverified or possible
  3. 3Gap detectionMass balance, missing QC, unverified entries, unknown onward distribution
  4. 4AI narrationAmazon Bedrock receives only the engine's facts
  5. 5Grounding checkAny identifier or number not in the facts → output discarded
  6. 6Authorised human reviewWritten justification; open gaps must be acknowledged

Live demonstration Fictional data

Pick an ingredient lot. See where it went.

This runs the real trace engine against a small fictional plant: a seasoning supplier has notified a possible issue with smoked paprika. Try LOT-PAP-2409-117 first, then compare a lot with no issues.

Choose a lot and run a trace.

Manufacturer use cases

Built for plants where one ingredient touches many products

Snack & savoury manufacturers

Seasoning premixes fan a single spice lot into dozens of SKUs. Trace through intermediates automatically and see which packed lots actually shipped.

Multi-level BOMs

Spice & seasoning blenders

Respond to upstream notifications with a documented list of your own customers, lots and quantities — and the records still missing.

Supplier notifications

Dairy & ingredient processors

Link allergen-bearing powders to every downstream batch and confirm which certificates and inspections are on file for each.

Allergen control

Co-packers & contract manufacturers

Give brand owners a clear, evidence-backed scope for their products without exposing other clients' data.

Brand-owner reporting

Bakery & prepared foods

Short shelf lives mean speed matters. Combine stock-on-hand and delivery records to prioritise what to hold first.

Short shelf life

Mock recalls & audits

Run timed trace exercises on historic lots and keep the report and audit log as evidence of the exercise.

Readiness exercises

Architecture

Runs in your AWS region

Authoritative traceability relationships live in PostgreSQL (Amazon RDS / Aurora). Documents are stored in Amazon S3, read by Amazon Textract, interpreted by Amazon Bedrock, and processed in the background by AWS Lambda.

CRM workflows for prospects and onboarding run through HubSpot; production data never leaves your environment for the CRM.

  • PostgreSQLLots, batches, consumption edges, packing runs, deliveries, counts
  • Amazon S3CoAs, inspection sheets, supplier letters, BOLs — versioned & encrypted
  • Amazon TextractForms & tables from scanned and PDF documents
  • Amazon BedrockField interpretation and investigation-summary wording
  • AWS LambdaIngestion, extraction, nightly mass-balance checks, report drafts

Pricing

Priced per production site

Illustrative pricing for this demonstration. Annual billing; implementation quoted separately.

Pilot

$1,450/ site / month

Prove the trace on one site with historical data.

  • 1 production site
  • Up to 10 users
  • CSV & SFTP data import
  • Relationship explorer & reports
  • Document extraction — 500 pages/month
Start with a pilot

Most chosen

Plant

$3,200/ site / month

Live integration for day-to-day investigations.

  • Unlimited users per site
  • MES / ERP / WMS connectors
  • Traceability Assistant
  • Incident workspace & review workflow
  • Document extraction — 5,000 pages/month
  • Quarterly mock-recall support
Talk to sales

Enterprise

Custom

Multi-site groups and co-manufacturing networks.

  • Cross-site tracing
  • Deployment in your AWS account
  • SSO / SAML & custom roles
  • Brand-owner portals
  • Dedicated success manager
Contact us

Request a demo

See it on your own product structure

We'll walk through a trace on a product structure like yours and discuss what data you already capture. Typical session: 45 minutes.

  • Bring a recent supplier notification or mock recall scenario
  • No production data required for the first session
  • Technical deep-dive on the AWS architecture available

FAQ

Frequently asked questions

Does the AI decide which products are affected?

No. Affected batches, lots and deliveries are computed by a deterministic trace engine from your recorded relationships. AI is used to extract document fields and to word summaries, and its text is rejected if it mentions any record or number that is not in the engine's output.

What happens when records are missing or unclear?

They are shown as gaps with a severity: unaccounted quantities from mass balance, manual entries awaiting verification, illegible line-log entries, missing inspection records and distributors whose onward shipments are unknown. Ambiguous records are listed against every candidate lot and shown as "possible", never silently resolved.

Can BatchBeacon initiate a recall?

No. It prepares scope and documentation. Only a user with the authorised-reviewer role can record a scope decision, and doing so requires a written justification and explicit acknowledgement of any open high-severity gaps. Whether and how to recall is your organisation's decision.

Is BatchBeacon certified for regulatory compliance?

BatchBeacon does not hold or confer any regulatory certification, and using it does not by itself demonstrate compliance with any regulation or scheme. It helps you organise and evidence the traceability records you keep.

Which systems does it connect to?

Common sources are ERP/MRP (lots, receipts), MES or line terminals (consumption), WMS (stock and shipments) and document stores. The Pilot plan starts with scheduled CSV or SFTP exports.

Where is our data stored?

In PostgreSQL and Amazon S3 in the AWS region you choose, encrypted at rest. On the Enterprise plan it can run in your own AWS account. Bedrock calls receive only the minimum facts needed for a summary.

Is the data on this website real?

No. Every supplier, product, lot, customer and document here is fictional and exists to demonstrate the software.