Software Development for Manufacturing and Industrial MES, IIoT & OEE
MES, ERP integration, IIoT and edge ingestion, OEE dashboards, predictive maintenance, quality management, and traceability platforms. ISA-95 aligned. Shipped in 12 to 24 weeks. USD pricing.
We map your build against ISA-95, your existing ERP, and your shop-floor data sources before we propose a stack.
Get started in 60 seconds
Who we've built for.
How we work in manufacturing
- What we build
- MES · IIoT · OEE · Predictive maintenance · Quality · Traceability · ERP extension · Digital twin · Andon
- Stack
- Next.js · Node.js · Python · PostgreSQL · TimescaleDB · AWS IoT · Azure IoT · OPC UA · MQTT · Kafka
- Compliance
- ISA-95 · ISA-88 · ISO 9001 · ISO 14001 · FDA 21 CFR Part 11 · GAMP 5 · IATF 16949 · AS9100
- Integrations
- SAP · Oracle · Microsoft Dynamics 365 · Infor · Plex · Ignition · Wonderware · OPC UA · MQTT · Modbus
- Pricing in USD
- Starter build from $21,000 · MES module from $28,000 · Predictive maintenance platform from $32,000
- Output
- Production platform · shop-floor integrations · runbook · on-call coverage · SLA reporting
Manufacturing software is where IT and OT meet, and where most projects either bridge or fall into that gap. The shop-floor data is real-time, brittle, and rarely documented. The corporate ERP is structured, governed, and slow to change. The work is making both layers play. This page covers what makes manufacturing builds different, the segments we serve, what ISA-95 actually means at the data layer, the named clients we have shipped for, and what every category typically costs.
Named manufacturing builds
Recent manufacturing and industrial builds.
Industrial-adjacent reseller workflow with metering integration and audit logging.
Read case study →Production monitoring platform with sensor ingestion and operator dashboards.
Read case study →Industrial yard management integrating with shop-floor logistics workflows.
Read case study →Manufacturing segments we serve
Manufacturing Execution Systems (MES)
Production order execution, dispatch, recipe and BOM, operator UI, OEE capture, downtime tracking, scrap and rework. ISA-95 Level 3 compliant.
IIoT and edge ingestion
PLC and SCADA data ingestion via OPC UA, MQTT, Modbus, EtherNet/IP. Edge gateways (AWS IoT Greengrass, Azure IoT Edge, Ignition). Time-series storage in TimescaleDB or InfluxDB.
OEE dashboards and shop-floor reporting
Real-time OEE (availability, performance, quality), downtime analysis, Pareto charts, shift comparison, machine ranking. TV-display dashboards on the floor.
Predictive maintenance
Vibration, temperature, current, and acoustic sensor ingestion. ML models for anomaly detection and remaining-useful-life prediction. CMMS integration (Maximo, Fiix, UpKeep, eMaint).
Quality management
SPC, FMEA, control plans, non-conformance tracking, CAPA workflow. Integration with quality test stations and metrology.
Traceability and genealogy
Lot, batch, and serial tracking from raw material through finished good. Forward and backward genealogy. Recall readiness. GS1 EPCIS compliant.
ERP and PLM extension
Custom workflows on SAP, Oracle, Microsoft Dynamics, Infor, Plex. PLM extension on Windchill, Teamcenter, Aras, Arena.
Digital twin and simulation
Process twin for what-if analysis. Asset twin for predictive maintenance. Built on Azure Digital Twins, AWS IoT TwinMaker, or custom.
Related services: Custom Software Development, AI & Machine Learning, Automation Platforms, API Integration, Cloud & DevOps, Data Pipeline Engineering, WMS Development.
Use cases — concrete examples with cost ranges
MES module v1
Production order dispatch, operator UI for run, OEE capture (availability, performance, quality), downtime reason capture, scrap and rework tracking, end-of-shift report. Integration with ERP for order pull and completion push. Stack: Next.js plus Node.js plus PostgreSQL plus TimescaleDB. Typical build 14 to 20 weeks. From $28,000 depending on line count and ERP integration depth.
IIoT and OEE platform
Edge ingestion from PLCs and SCADA via OPC UA or MQTT. Real-time OEE dashboards. Downtime analysis with Pareto. Shift comparison and machine ranking. TV-display dashboards on the floor. Stack: Ignition or AWS IoT Greengrass edge plus TimescaleDB plus Grafana or custom dashboards. Typical build 12 to 16 weeks. From $20,000 depending on line count and protocol mix.
Predictive maintenance platform
Sensor ingestion (vibration, temperature, current, acoustic). Anomaly detection model. Remaining-useful-life prediction. CMMS integration for work-order auto-creation. Operator alert UI. Stack: Python ingestion plus TimescaleDB plus MLflow plus React UI. Typical build 16 to 24 weeks. From $32,000 depending on asset count and model complexity.
Traceability and genealogy
Lot, batch, and serial tracking from raw material through finished good. Forward and backward genealogy queries. Recall readiness. GS1 EPCIS compliant. Integration with WMS, MES, and ERP. Typical build 12 to 18 weeks. From $25,000 depending on product complexity and existing system integration.
Why manufacturing software is different
Manufacturing builds carry three costs that a standard SaaS does not. First, IT-OT integration: every PLC, every SCADA, every legacy MES speaks its own protocol (Modbus, OPC UA, MQTT, proprietary serial). Bridging shop-floor data into modern data platforms is its own engineering programme. Second, data fidelity and timestamps: when you are calculating OEE, late or duplicate sensor readings invalidate the metric. The data pipeline has to handle out-of-order events, clock drift, and downtime gaps without lying. Third, change management on the floor: shop-floor users are not patient. A new UI that adds two clicks to a quality check costs production minutes that costs money. Builds succeed or fail on operator buy-in. We design manufacturing builds around three principles. Establish a single source of truth at the ISA-95 Level 3 boundary so MES, ERP, and shop-floor reporting agree. Treat shop-floor data as time-series first, transactional second, because OEE, predictive maintenance, and downtime analysis depend on millisecond-grained event streams. Co-design with operators from week 1 because adoption, not architecture, is the most common failure mode.
Implementation roadmap
Five-phase rhythm for manufacturing builds. Operator co-design runs from week 1.
Discovery and IT-OT mapping (2 weeks)
Shop-floor walk-through. PLC, SCADA, and existing MES inventory. ERP integration scope. Operator interview round. Output: data-source inventory plus IT-OT integration map.
Architecture and data-model design (2 weeks)
ISA-95 level boundaries. Time-series schema. Event model for OEE and downtime. ERP integration pattern. Output: ADRs plus data dictionary.
Build (10 to 18 weeks)
Two-week sprints. Operator UI co-design every sprint with a shop-floor user group. Edge ingestion built in parallel by IIoT squad.
Plant validation and UAT (2 weeks)
Pilot line deployment. OEE numbers reconciled against existing tracking. Operator training. Edge cases tested (network drop, sensor failure, ERP downtime).
Plant rollout and dual on-call (2 weeks initial line plus phased rollout)
Production deploy on pilot line. Monitoring. On-call rotation. Phased rollout to additional lines or plants on a 2 to 4 week cadence.
Tech stack and architecture
Default manufacturing stack. Each layer chosen for IT-OT bridging, time-series throughput, and operator-friendly UX.
- Operator UI: Next.js with TypeScript for shop-floor screens. Optimised for tablet and touchscreen. Offline-tolerant. WCAG 2.2 AA where regulatory or contractual.
- Application layer: Node.js or Python (FastAPI) for MES and reporting services. Stateless behind API gateway. Event-driven for OEE and downtime aggregation.
- Edge ingestion: AWS IoT Greengrass, Azure IoT Edge, or Inductive Automation Ignition as edge gateway. OPC UA, MQTT, Modbus, EtherNet/IP, Sparkplug B as supported protocols.
- Time-series: TimescaleDB or InfluxDB for sensor readings, OEE events, downtime. PostgreSQL for transactional MES state.
- ML and analytics: Python with scikit-learn, PyTorch, or LightGBM for predictive maintenance and anomaly detection. MLflow for model versioning. Feature store for reuse.
- ERP and PLM integration: SAP IDoc, BAPI, or SAP API Business Hub. Oracle E-Business Suite or Fusion. Microsoft Dynamics 365 via Dataverse. PLM via API or file-based exchange.
- Cloud and edge: AWS or Azure for cloud workloads. Edge gateways at the plant for offline tolerance. Bidirectional sync with conflict resolution.
Compliance and audit readiness
Every manufacturing build ships with the controls and documentation auditors and customers expect. ISA-95 for IT-OT functional boundaries. ISA-88 for batch process control. ISO 9001 for quality management. ISO 14001 for environmental management. FDA 21 CFR Part 11 plus GAMP 5 for regulated pharmaceutical or medical-device manufacturing. IATF 16949 for automotive. AS9100 for aerospace. Builds in regulated industries ship with electronic signature, audit trail, and validation documentation that supports your QA team or notified body submission.
Cost drivers we see in manufacturing builds
Manufacturing build cost ranges widely. These are the seven drivers we see push or pull on the number.
- PLC and SCADA protocol mix. Single-protocol (OPC UA only) is fast. Mixed (OPC UA plus legacy Modbus plus proprietary serial) doubles edge integration cost.
- Line and asset count. Single line is straightforward. Multi-plant with 100+ lines is its own rollout programme over 6 to 18 months.
- ERP integration depth. Read-only is cheap. Bidirectional with order completion, inventory move, and quality result push adds 25 to 40 percent.
- Regulatory scope. Non-regulated is the baseline. FDA 21 CFR Part 11 plus GAMP 5 validation adds 40 to 60 percent and 4 to 8 weeks of validation work.
- ML and predictive maintenance maturity. Off-the-shelf anomaly detection is fast. Custom remaining-useful-life models per asset class add months and substantial data engineering.
- Operator UI co-design depth. Cookie-cutter UI is cheap and rarely adopted. Co-designed UI with shop-floor user research adds 15 to 25 percent and dramatically improves adoption.
- Multi-plant rollout. Single plant is one project. Multi-plant rollout adds change management, training, and per-plant variation that easily adds 50 percent.
Pricing
Manufacturing MVP
From $21,000
- Single line, single workflow (OEE or downtime tracking), ERP read-only.
- 12 to 16 weeks.
MES module
From $28,000
- Single plant, multiple lines, ERP bidirectional, operator UI co-design.
- 16 to 22 weeks.
IIoT and analytics platform
From $27,000
- Edge ingestion, time-series storage, dashboards, no MES.
- 12 to 16 weeks.
Predictive maintenance platform
From $32,000
- Sensor ingestion, ML models, CMMS integration.
- 16 to 22 weeks.
Traceability platform
From $25,000
- Lot, batch, serial tracking with WMS and MES integration.
- 12 to 18 weeks.
Maintenance retainer
From $4,500 / month
- On-call cover, edge gateway monitoring, ERP integration tracking, ML model drift monitoring. SLA-backed.
Manufacturing trends shaping 2026 builds
Six manufacturing-software shifts shaping 2026 builds.
- ISA-95 still the governing pattern. Most net-new MES builds align to ISA-95 Levels 1 through 4 to enable IT-OT interoperability and future-proof against ERP and shop-floor changes.
- Edge AI moving to production. Vision inspection at the line via NVIDIA Jetson or AWS Panorama. Anomaly detection running on-device at the gateway.
- Generative AI for operator support. RAG over SOPs, work instructions, and machine manuals. Operator co-pilot for troubleshooting and maintenance procedures.
- Sustainability and ESG reporting. Scope 1, 2, and 3 emissions tracking integrated with production data. CSRD in EU, SEC climate rule in US drive this into the MES scope.
- Digital twin maturing past the pilot stage. Process twins for what-if analysis and asset twins for predictive maintenance moving into production deployments.
- Composable MES vs. monolithic vendors. Shift from monolithic MES (SAP DMC, Plex, AVEVA) to composable approaches built on event streams and ISA-95 boundaries.
FAQ
Yes. OPC UA is our default for modern PLC and SCADA integration. MQTT (and Sparkplug B) for IIoT and lightweight ingestion. Modbus, EtherNet/IP, and proprietary serial supported through edge gateways (Ignition, AWS IoT Greengrass, Azure IoT Edge).
Yes. SAP via IDoc, BAPI, SAP API Business Hub, or Plant Connectivity. Oracle E-Business Suite or Fusion via REST API. Microsoft Dynamics 365 via Dataverse and Power Platform. Infor, Plex, IFS supported through API or file-based exchange.
ISA-95 Levels 1 through 4 boundaries respected by default. Level 3 MES is where most of our work sits. Level 4 ERP and Level 2 SCADA integrations defined at the start of every project to avoid scope creep.
Yes. Sensor ingestion (vibration, temperature, current, acoustic). Anomaly detection and remaining-useful-life prediction models trained on historical asset data. CMMS integration for work-order auto-creation. We deploy with model versioning and drift monitoring.
Yes. Electronic signature, audit trail, validation documentation. GAMP 5 categorisation applied per build. We do not file the validation submission ourselves. We deliver the validation deliverables to your QA or notified body.
Yes. Phased rollout with per-plant configuration. Change management materials, operator training, and runbook delivered per plant. Typical rollout cadence is 2 to 4 weeks per additional plant after pilot.
Yes. NVIDIA Jetson or AWS Panorama for vision. Custom CNN models trained on inspection samples. MLOps wired into the edge so model updates roll out under governance.
SPC charts with auto-detect rule violations. FMEA tracking. Control plans. Non-conformance and CAPA workflows. Integration with quality test stations and metrology equipment.
Yes. API or file-based exchange depending on the PLM version. Engineering change orders, BOM sync, and document management. Most PLM integrations also touch the ERP.
Yes. Azure Digital Twins, AWS IoT TwinMaker, or custom. Process twins for what-if analysis and capacity planning. Asset twins for predictive maintenance. We start with the business outcome and scope the twin to it, not the other way around.