Bring a new plant online — without a device list
Create the plant, connect the VPN to its router — Mirox does the inventory work. The on-site agent rolls out by itself, a network scan finds every reachable device, and your uploaded documents fill in the master data field by field, for you to accept or correct.
Commissioning usually means typing in every device
Classic monitoring rollouts start with a list: every address, every inverter, every datalogger entered by hand — often only after a site visit, and again whenever the hardware changes. The list is out of date before it is finished, and every mistake in it shows up later as a gap in the data.
From plant record to live data
A guided checklist walks each step and unlocks the next one once it is done.
Create the plant
Portfolio, type, name, DC peak power, date of build and the exact coordinates — the coordinates are what the weather model and the agent configuration need. On save the plant is activated and its agent rollout starts in the background.
Upload the documents
Yield reports, datasheets, contracts. Every file you add is analyzed and tagged automatically, so the plant arrives with its paperwork already sorted.
Let the AI propose the master data
The platform reads the uploaded documents and proposes values field by field — basics, address, dates, components, key contacts. You accept, edit or reject each proposal.
Connect the VPN to the router
One VPN connection links the platform to the plant router, and the on-site agent is redeployed with it automatically. Once the connection reports live, the platform can reach the plant’s dataloggers and inverters.
Scan, then adopt
Mirox scans the local network behind the VPN and lists every reachable device. Map each one to a logger adapter — from then on production data, alarms and status flow in.
The demo above is the real onboarding guide
Everything in it runs on demonstration data — no login, nothing to install, and nothing you do touches a real plant. The plant you land on is mid-onboarding: walk its checklist from the first step, let the analysis fill the record, connect the VPN, run the discovery scan and adopt the loggers — until it reports live data on the dashboard. Every other screen of the platform is open to you from there.
What comes online by itself
Once the VPN is up, the inventory is discovered rather than typed — and where a device is unknown, there is a path for it too.
Every reachable device
The scan works through the plant network behind the VPN and lists what answers — routers, dataloggers, inverters, meters, switches and cameras — with vendor, address and reachability. The scan ranges follow the plant’s VPN configuration; you can narrow a run to a sub-range and stop it at any time.
Loggers adopted, not configured
Each discovered device is mapped to a logger adapter in a guided step. From then on the platform collects production data, alarms and status from it. For a maintenance window you can pause a logger — from fifteen minutes up to a day — and resume it afterwards, without the downtime raising false alarms.
Components discovered automatically
Behind a connected logger the component hierarchy is discovered on its own — inverters and meters, and for batteries down to the individual cell — feeding the standardized metric set.
Agent rollout without a wizard
You never deploy anything yourself. Creating the plant triggers the agent rollout, and every VPN change redeploys it automatically — a status banner shows whether it is deploying, updating or not responding.
A logger we have never seen
Not every logger has a ready-made adapter. A mapping wizard lets you bind its raw values to the standardized metrics one by one, with a dry run against the live device before you save.
Adapters for anything reachable
If a device needs its own adapter, we build one against whatever interface it offers — REST, a web interface, a database or files — including undocumented legacy equipment.
Your documents become the plant record
Point the wizard at a yield report and it proposes the entire plant record — basics, address, dates, business data, components, contacts. Nothing is written silently: every proposal is an editable field with an accept or a reject next to it.
- Register a new plant straight from a yield report
- Audit an existing plant against a fresh document and see what differs
- Re-read the already uploaded documents and propose updates at any time
- Every proposed value stays editable before you accept it — the AI never writes to your master data on its own
Frequently asked questions
Do I have to enter my devices manually?
No. Once the VPN to the plant router is up, Mirox scans the local network and lists every reachable device; you map each one to a logger adapter and the data starts flowing.
Who installs the on-site agent?
Nobody on your side. Creating the plant triggers the rollout in the background, and every VPN change redeploys the agent automatically — a status banner shows whether it is deploying, updating or not responding.
What does the AI change on its own?
Nothing. It proposes values read from your documents, each one an editable field with an accept or reject — only what you accept is written to the plant record.
Is my hardware supported?
Around twenty vendor adapters today (SMA, Sungrow, Huawei, Janitza, Phoenix Contact, Bluelog, NR Electric and others) with 450+ standardized metric types, and a custom adapter for new hardware built on request — typically within days.
Ready to Simplify Your Operations?
See how Mirox can bring your team together and save you money. Free pilot available — connect your first plant at no cost.