OEE Monitoring
Availability, performance, and quality streamed from the line in real time — with loss trees that point to the exact station eating your capacity.
Febriansyah Tech System is an AI-powered manufacturing intelligence platform that connects factory PLC and machine data with Claude. It turns real-time production data into insight for OEE monitoring, anomaly detection, production analysis and machine performance — and lets your production team ask questions in plain language.
Cycle time drifted +14% over the last 9 batches. Matches a tooling-wear signature last seen 41 days ago.
“Shift 2 lost 12 min at Press-02 to micro-stops — 68% of the shift’s shortfall.”
Production data already exists in PLCs, SCADA and historians — but it reaches people as dashboards nobody opens or spreadsheets compiled after the shift is already lost.
Data-rich, insight-poor
Ask, and get an explanation
Claude sits on top of a manufacturing data model, so each module shares one version of the truth about your line.
Availability, performance, and quality streamed from the line in real time — with loss trees that point to the exact station eating your capacity.
Statistical baselines plus Claude reasoning flag drift, micro-stops, and out-of-tolerance cycles before they become scrap or downtime.
Shift, batch, recipe, and SKU-level breakdowns generated on request. Ask for last night's shortfall and get the contributing causes.
Cycle time, throughput, and utilisation per asset — benchmarked against its own history so degradation is visible, not assumed.
A four-stage pipeline that keeps industrial data trustworthy all the way to the reasoning layer.
Edge collectors read PLC registers and machine telemetry on the plant floor. Read-only, no changes to control logic.
Raw tags are normalised into a manufacturing ontology: assets, lines, shifts, recipes, and units of measure.
Claude works over retrieved production context — metrics, events, maintenance notes — through RAG and agentic tool use.
Answers come back as language your production team can act on, with the numbers and time windows they were derived from.
Independent, bootstrapped software company. Founded and engineered by a developer with 9+ years in manufacturing IT — not a theory-first AI project.
9+ years building production systems for manufacturing — real-time OEE telemetry on automotive assembly lines, MES/LES support for FMCG plants, and SAP ERP integrations at enterprise scale.
Built a real-time OEE telemetry dashboard on the factory floor using Node.js, WebSockets, Node-RED and industrial PLC hardware — cutting downtime-tracking latency from hours to sub-second.
Sustained 99.9% uptime for Logistic & Manufacturing Execution Systems (LES/MES) across manufacturing facilities, resolving critical production incidents.
Architected an SAP ERP integration portal that replaced external vendor contracts, saving 800 million IDR per year.
This is why the platform starts with PLC reality and production constraints — not with a model demo.
We want to develop and validate the platform against real manufacturing data — and we need both runway on the Claude API and engineering depth to get the architecture right.
Credits let us run the platform against continuous, high-volume factory telemetry during pilots — measuring latency, cost per query, and reasoning quality on real production data instead of synthetic samples.
This support helps us validate the product with factories and scale it into a production SaaS platform.
Wire live PLC and machine feeds from partner factories into the platform and prove the analytics against ground-truth production records.
Seeking pilot factoriesRun structured pilots with production teams, harden the anomaly models, and build the evaluation sets that make results defensible.
Design partnersMulti-tenant onboarding, connector library, and predictable unit economics — a production SaaS platform for manufacturing intelligence.
Commercial releaseIf you run a plant and want to pilot the platform on real production data — or you’re on the Claude team and can support us with credits and technical guidance — we’d like to talk.