Connect any PLC. Store every tag. Act on the answer.
KMS is the manufacturing intelligence platform we built because no off-the-shelf historian could keep up with what our customers were asking for. Microservice-native, fully no-code, with a soft gateway that centralises data from every site — and an action engine that reasons over PLC signals and video streams.
One stack, six layers, no glue code.
From the field gateway up to the LLM, KMS is a single product. You don't integrate four vendors to get one answer.
From every site, every camera, every PLC — into one platform.
Soft Gateway sits at each plant, reads the fieldbus, buffers if the link drops, and forwards to the central KMS cluster. Cameras stream alongside tags into the same model. Microservices on the right do the work.
Pure microservices. Pure no-code.
Every capability in KMS is an independently deployable service — scale only what your plant needs, replace any piece without taking the rest down. And every workflow is built by dragging blocks, not writing code.
Vision AI · no-helmet
zone = packing
halt conveyor + SMS
— drag blocks · publish · live in < 30 seconds —
Everything a manufacturing intelligence stack needs to do.
A PLC that reads its own past — and refuses to repeat it.
KMS is the only platform that ships a Soft PLC with first-class access to the historian. Decision logic can read minutes, hours or months of context inside the rung — preventing duplicate batch starts, enforcing golden-batch envelopes at runtime, blocking retries that have already failed, and arbitrating across sites that a wired PLC cannot see. No other PLC platform on the market reads its own historian inside the logic.
- Live tags and full historical context as first-class inputs to ladder / structured text
- Block duplicate batch starts within configurable look-back windows — “has B-1042 run in the last 24h?”
- Golden-batch envelope enforcement at runtime — abort on drift, before quality is lost
- Replay any incident window through the same rung to validate logic before deploying to the wire
- Maker-checker write-back to physical PLC, fully audited (Part 11 ready)
One soft gateway, every site, never loses a tag.
Lightweight, deploys on any Linux box or industrial PC at the plant. Reads the fieldbus, normalises into the KMS model, and forwards to central. Pair it with a standby gateway and a flaky WAN becomes a non-event.
- Active/standby pair with sub-2-second failover
- Local time-series buffer up to 30 days when the WAN is down
- Chronological back-fill with deduplication on reconnect
- Single binary; managed centrally — no per-site config drift
Cameras are first-class signals.
RTSP streams flow into KMS the same way PLC tags do. Vision microservices run PPE detection, perimeter protection, restricted-zone intrusion and people counting — and fire the same action engine you already use for tags.
- PPE detection — helmet, vest, goggles, gloves
- Perimeter & restricted-zone breach alerts
- People counting & zone-occupancy
- Frames cross-correlated with tag history for incident review
Native drivers, not adapters.
We've written our own connectors for the protocols our customers use. Whether you're running Siemens S7, Rockwell ControlLogix, Schneider M580, ABB AC800 or a Yokogawa DCS — KMS reads it natively.
- Auto-discovery of OPC UA address space
- Tag mapping with engineering units & deadbands
- Edge gateway for store-and-forward over flaky links
- Bi-directional — read for history, write for control
Built for tag history at plant scale.
Purpose-built time-series core handles millions of tags with second-level granularity. Hot, warm and cold tiers; configurable retention; pre-aggregated rollups so dashboards stay fast even on a year of history.
- Compression tuned for industrial signal patterns
- Per-tag retention & downsampling policies
- Continuous queries for OEE and rate calculations
- Replication across edge ↔ central instance
Drag, drop, ship to the line.
Operators get role-based dashboards on shopfloor terminals or mobile. Plant managers get OEE, downtime reasons and SKU-level production reports without exporting anything to Excel.
- Drag-and-drop dashboard builder, no code
- OEE with availability / performance / quality breakdown
- Alarm banner & live trend in one view
- Mobile shopfloor view for supervisors on the move
From insight to setpoint, automatically.
Rules and events trigger actions: write a setpoint back to the PLC, raise a ticket, page an operator, hand off to MES, or call any HTTP endpoint. Closed-loop control without a separate workflow tool.
- Visual rule builder + scripted rules for power users
- Audit trail on every action (Part 11 ready)
- Permission model with maker-checker for write-back
- Outbound channels: PLC, email, Teams, webhook, SMS
Ask the historian in plain English.
KMS connects your contextualised tag store to an LLM. “What caused yesterday's downtime on Line 2?” returns a grounded answer with the underlying trend and the alarm sequence to back it up.
- Retrieval over your tag history, alarm log & batch records
- Citations to the trend window that justifies each answer
- Self-hosted models supported for air-gapped sites
- No tag data leaves your perimeter
Speaks to whatever you already own.
Kickoff to go-live in weeks, not months.
Traditional SCADA modernisation runs 6–12 months — protocol stacks to write, dashboards to build from scratch, change requests for every alert. KMS ships with the protocols, the dashboards, and the workflow builder already in the box. Median deployment across our installed base sits at four weeks.
- Procure SCADA / historian licences
- Custom protocol drivers + tag mapping
- Dashboard builds per plant
- Change-request loop for every alert
- UAT + parallel run
- Week 1 · tag map + auth, soft-gateway live
- Week 2 · OEE / line dashboards in production
- Week 3 · workflows, alerts, action engine
- Week 4 · go-live + handover to ops
Protocols already in the box
Modbus, OPC-UA, IEC-60870, Profinet, BACnet, MQTT, S7 — native drivers. No firmware-side glue code on day one.
Templates per industry
Sugar, edible-oil, datacenter, EMS, batch — dashboards and OEE roll-ups land pre-wired. Configure, don't build.
No-code workflows
Your control engineers wire alerts, batch logic, recipes without a backend change request. Updates ship same-day.
Same artifact, anywhere
Air-gapped on-prem, hybrid, or cloud — identical container stack. No re-architecture between PoC and production.