Bridge the plant floor to the systems you already run.

One console. N site runtimes. Collect PLC tags on site, keep four hours of buffer locally, and publish to MQTT, Kafka, InfluxDB, PostgreSQL, or HTTP webhooks — without replacing your broker, historian, or UNS platform.

Built for industrial connectivity

OT collection. Site isolation. Your sinks, your contract.

Like a connectivity gateway, we sit between PLCs and IT — but we are shaped as a control plane plus partitioned site runtimes, not another all-in-one cloud stack.

  • Collect on site Poll PLCs with multi-group intervals, transforms, browse, audited write tags, and suspend/resume when maintenance starts.
  • Process at the edge Filter with isolated SQL rules before publish. Keep noise off the broker and the sinks.
  • Deliver northbound MQTT JSON or Sparkplug into EMQX, HiveMQ, or cloud brokers — plus Kafka, InfluxDB, PostgreSQL, and HTTP webhook sinks.

High-level architecture

Plant floor → site runtime → your broker & sinks.

One Console for config and audit. N partitioned site runtimes for collection, four-hour buffer, and publish. Catalog drivers to the PLCs you name; MQTT, Kafka, Influx, Postgres, and HTTP webhook Destinations to systems you already run.

Brand marks identify interoperability targets. Protocol maturity stays honest: Modbus TCP beta; other catalog drivers available until a named PLC passes conformance. BACnet, DNP3, IEC 104, and OPC DA are on the roadmap — not field-supported. Logos © their owners.

How it works

Your industrial data path, clearly split.

Console configures. Sites collect and publish. The broker remains yours — the same pattern industrial teams already know from edge gateways, with honest boundaries.

  1. 01
    Connect

    Point a site runtime at named PLCs. Start with Modbus TCP (beta) or an available catalog protocol for a scoped pilot. Test connection before you trust the line.

  2. 02
    Buffer

    Keep a four-hour local WAL on the site. Last-value retain and MQTT reconnect replay cover broker blips. This is an outage buffer — not a historian.

  3. 03
    Publish

    Deliver standardized tag payloads to MQTT topics — or Kafka, InfluxDB, PostgreSQL, and HTTP webhook Destinations. Optional ISA-95 prefix. Rules can fan filtered events, or write a tag on a rising edge.

Capabilities

What ships for a real plant pilot.

Every claim below is in the product today. Protocol support levels stay honest: beta and available until named PLC evidence says otherwise.

01

Plant-floor connectivity

Multi-group polling, tag browse (Modbus / OPC UA), CSV import, scale/offset/deadband, connection test, audited PLC write tags, and suspend for maintenance. Catalog lists many engines; Modbus TCP is beta — others are available for named pilots.

02

Reliable edge buffering

Process-local store-and-forward with four-hour retention. When the broker returns, sites replay. Design point is 5,000 messages/s per site — an engineering target, not a published SLA.

03

Northbound applications

Console Destinations: MQTT / MQTTS JSON and Sparkplug B into EMQX, HiveMQ, Mosquitto, or cloud brokers (AWS / Azure / GCP). Hosted sinks also include Kafka, InfluxDB 2.x, PostgreSQL, and HTTP webhooks (HMAC, retry, TLS). Isolated rules: SQL filter or passthrough — optional rising-edge PLC write.

04

Operator-grade control plane

Local users with viewer / operator / engineer / admin RBAC, optional LDAP, secrets encrypted at rest, Console HTTPS / site mTLS, config export/import, admin audit CSV, SMTP and webhook alerts, Grafana + Prometheus, site enrollment tokens.

Where it fits

When MakeItIntelligent is the right move.

We win when the plant already has a broker and needs honest site runtimes — not another platform that owns dashboards, UNS modeling, and SSO in one SKU.

  • Line & site connectivity Stand up one site runtime per plant area. Isolate PLCs by edge node so two sites never share the same controller.
  • SCADA / MES modernization Publish tags to a central MQTT tree. Decouple devices from applications the way modern industrial MQTT architectures intend.
  • Broker-first IT teams Keep EMQX or HiveMQ as the system of record for topics. Add RBAC and audit on the control plane without buying a full UNS suite.

First pilot

One site. One named PLC family. Live tags on your topic.

Success is not a protocol-count slide. It is tags flowing and one rule publishing to a topic you already own.

  1. Stand up Console and one site runtime on the management network.
  2. Connect a real PLC — prefer Modbus TCP beta, or name vendor/model/firmware for an available driver.
  3. Publish to the customer broker or sink (MQTT / Kafka / Influx / Postgres / HTTP) with TLS if required.
  4. Prove reconnect buffer, suspend/resume, and operator vs engineer roles with your team.

FAQ

Straight answers.

Is this another all-in-one industrial IoT platform?

No. MakeItIntelligent is a control plane plus partitioned site runtimes: collect on site, buffer locally, publish into the broker and sinks you already run. We do not replace your historian, dashboards, or UNS platform.

How many protocols are production-ready?

Modbus TCP is beta. 30 catalog entries are selectable; BACnet, DNP3, IEC 104, and OPC DA are roadmap-only. Other drivers stay available until a named physical PLC passes our conformance gates. We do not claim 100+ production drivers.

Where does historical data live?

Not here. The site bus keeps about four hours locally. Long-term storage stays in your MQTT/MES/cloud stack.

What if Console is down?

Site runtimes keep the last desired state and continue polling and publishing. Console is configuration and audit — not the hot path.

Can you write back to the PLC?

Yes, in a narrow model: tags with a write attribute (audited Console write) and filter-rule rising-edge writes to the same device. Not a control loop or a safety function.

Sparkplug, Kafka, Influx, Postgres, HTTP?

Hosted northbound today: MQTT/MQTTS JSON, Sparkplug B (not TCK-certified), Kafka, InfluxDB 2.x, PostgreSQL, and HTTP webhooks (HMAC + retry). These publish; they are not a product historian.

LDAP, HTTPS, alerts?

Optional LDAP bind (group→role). Optional Console HTTPS and site mTLS. Secrets sealed at rest. SMTP and webhook alerts for driver, Destination, queue, and edge-offline. OIDC/SAML SSO is not in this release.

Start connecting the plant floor — honestly.

Tell us the site, the PLC family, and the broker you already run. We will scope a pilot that fits your network — not a slide deck full of drivers you will never certify.

We do not claim certified Sparkplug, unbounded PLC control, HA failover, or 100+ production protocols.