About IONECT

We connect things that didn't talk before.

IONECT is an engineering company. We build the hardware, software and integration that lets industrial machines report, explain and converse — starting with EV charging, expanding to the wider IoT fleet.

Who we are

A small team of engineers. We have spent the last seven years building real machine systems — first IoT prototypes, then EV charging infrastructure at scale, and now the language layer that sits on top of both.

We work across the full stack: silicon, firmware, OCPP and other industrial protocols, cloud, and language models. Few teams work end-to-end like this. It is the reason our products survive contact with real operators.

Three eras, one company

The work has changed shape three times. The question has not.

2014 — 2017
Era 1 · IoT

Sensors to dashboards.

Embedded prototypes, LoRa and MQTT networks, custom telemetry for partners who needed visibility into machines they couldn't see. The foundation we still build on.

2018 — 2024
Era 2 · EV

Chargers to the grid.

EV charging infrastructure at scale — chargers, OCPP, cloud, payments and fleet logic. We did not study the operator's problem; we lived inside it.

2025 — now
Era 3 · Talkable

Machines to humans.

The same machines, with a voice. Language models that understand the protocols and data we have been working with for a decade. Dialogue, not dashboards.

How we work

Three principles
we don't negotiate on.

01

Start at the edge.

The interesting problems are in the sensors and the protocols. If the data going in is wrong, no language model on the way out will save it. We build from silicon up.

02

Earn the right to speak.

A talkable machine is not a chatbot. It says something only when it has something useful to say — a real anomaly, a real recommendation, a real change. Silence is a feature.

03

Build for operators, not decks.

We measure ourselves on whether the people running the machines actually open the thing again tomorrow. That is a harder bar than a polished demo, and the only one that matters.

Have a machine that should learn to talk?

Tell us what it does and who runs it. We will come back with a candidate stack and a plan.