01 — The plant, unified
Built for how
the plant
actually runs
One platform for MRP, procurement, inventory, quality and shop-floor analytics — running on the sensor data your equipment already speaks (OPC-UA, MQTT), not a patchwork of spreadsheets and point tools.
We reply within one business day. · Ingest: OPC-UA · MQTT · SAP OData
02 — The current stack is duct tape
The MRP stack
is broken. We built
one system in
its place
The average plant runs a legacy ERP for finance, a bolted-on MES for downtime, a separate scheduler for procurement, a BI tool over spreadsheets for anything a director actually asks about, and a Sharepoint for the rest. Every reconciliation between them is a place data goes stale. EvoWerk replaces the seams with one system.
Unified, not stitched
One system replaces the MES + ERP + PM + BI + spreadsheet stack. Every module reads and writes the same data model.
BOM · MRP · POs · lots · OEE · quality — one database, one audit trail.
Runs in your cloud
Customer-hosted by design. Your production data never leaves your AWS or Azure tenancy, and EvoWerk is never in your uptime path.
Terraform modules for ECS + RDS (AWS) and Container Apps + Key Vault (Azure).
Secure by construction
Multi-tenant isolation, MFA, and an append-only audit log — enforced at the database, not by app-code discipline.
Postgres trigger blocks UPDATE and DELETE on audit rows. Verified in CI on every push.
03 — No logos yet, here's the proof instead
Credibility,
measured in
mechanics
We haven't shipped the first customer yet, so we won't paint a logo wall we haven't earned. What we can show is what's actually in the box: the depth of the platform, and how tightly its security posture is enforced.
- 53test modules
Backend coverage across auth, MRP explosion, procurement, inventory, forecasting, RCA, and cross-tenant isolation.
- 0critical CVEs
On every push: SAST (bandit), pip-audit, npm audit, Trivy container scans. The build fails otherwise.
- 1DB-level trigger
Blocks UPDATE and DELETE on audit rows regardless of what the application code tries. Verified in CI against real Postgres.
- 2clouds, one Terraform
AWS (ECS + ALB + RDS) and Azure (Container Apps + Key Vault). Customer-hosted, in your account.
04 — Four surfaces, one platform
Streamline
the operation
Aggregate, automate and analyze in one place. Every module writes to and reads from the same operational data model — no exports, no middleware, no "we'll integrate that later."
- System of record
MRP Core
Run the plant.
BOMs and recipes with revisioning, netted-vs-gross MRP explosion, a build calculator, and procurement that generates POs off the plan — MOQ rounding and preferred-supplier logic included.
- Revisioned BOMs · netted MRP · lot/FEFO consumption
- Auto-generated POs from demand + on-hand
- Concurrency-safe stock adjustments, full audit trail
- Data ingestion
Integrations Hub
Speak what your machines already speak.
OPC-UA, OPC-DA, MQTT, SQL Server, Oracle, SAP OData, OSIsoft PI historian, generic REST, and CSV/Excel — pulled into one operational data model, no middleware layer to babysit.
- OPC-UA · OPC-DA · MQTT for shop-floor telemetry
- SAP OData · Oracle · SQL Server · OSIsoft PI
- DAG-based workflow builder for scheduled pulls
- BI · self-serve
Analytics
The floor, in numbers you can bet a run on.
OEE and SPC control charts, downtime Pareto and MTTR, scrap trends, throughput — rolled up to a plant summary. A report builder and pivot/SQL query layer over any dataset you have loaded.
- OEE · SPC · Pareto · MTTR · shift comparison
- Custom dashboards, saved reports, AI-narrated summaries
- Predictive maintenance and AutoML with SHAP explainability
- AI copilot
Ask
Every planner gets an analyst.
A natural-language chat over your live operational data. Ask for a specific line's downtime yesterday, a supplier's on-time-in-full trend, or the parts driving next week's procurement — no SQL, no dashboard hunting.
- Grounded in your loaded datasets — never hallucinated numbers
- Every answer cites the query it ran
- Works on OEE, inventory, POs, quality events, forecast plans
05 — Every business function, one system
Modules that
cover the plant
Core covers the floor as it runs today. Premium runs ahead of it — predictive maintenance, forecasting, vision and AutoML on the same data model the core writes to.
- BOM & RecipesCore
- MRP PlanningCore
- Procurement & POsCore
- Inventory & LotsCore
- Demand ForecastingPremium
- OEE & SPCCore
- Downtime & MTTRCore
- Quality & RCACore
- Predictive MaintenancePremium
- Computer VisionPremium
- AutoMLPremium
- Workflow AutomationCore
06 — Vocabulary layer, industry-aware
Industry-focused,
without a fork
The core is industry-agnostic. A per-industry vocabulary layer renames it (scrap becomes spoilage, shrinkage or return-rate as needed) so the interface reads native, without a separate codebase per vertical.
Discrete Manufacturing
Every unit, every serial, every scrap event.
Serialized inventory, revisioned BOMs, and OEE per line — the base configuration Phase 1 was built around.
≥ 85%OEE target visible per line, per shiftFood & Perishables
Lots, FEFO, and spoilage where scrap used to be.
Vocabulary layer renames "scrap" to spoilage; FEFO consumption on every lot; expiry-aware demand forecasting.
FEFOFirst-expiry-first-out enforced on consumptionPharma & Regulated
The audit log a compliance team can actually rely on.
Append-only audit trail enforced at the Postgres trigger level. Every actor, action, and payload — insert-only, database-side.
0Rows that can be updated or deleted, everAutomotive & Assembly
Revision-aware BOMs where the engineering change is the norm.
BOM revisions with as-designed / as-built separation; MRP explosion respects the revision on the build order.
n-revBOMs tracked across every engineering changeDistribution & Retail
Phase 2 — the same forecasting engine, without the OEE surface.
Standalone inventory analytics: on-hand, demand-trend, shrinkage/return-rate — same forecasting and report builder, no shop-floor modules.
Phase 2Extending Phase 1's forecasting layer, without MES
07 — Twenty minutes, on your data
See it on your
own numbers
We'll load a subset of your BOMs, one line's OEE feed and a supplier list, then walk through a live MRP explosion and the POs it generates. No slides.
We reply within one business day.