Home / For your engineers For your engineers

How it works, what it needs, what it costs to run

The detail your controls, IT and security people will ask about. The administrator’s manual has the rest.

How it works

Playbooks, the model of your plant, how a finding is made and where the results go.

How it works →

Anomaly Watch

The checks, the baselines per running state and product, and how a confirmed anomaly becomes an alarm you import.

Anomaly Watch →

Alarm Health

The six kinds of alarm problem, the thresholds, the proposed changes and the KPIs.

Alarm Health →

Guided setup

The prerequisites, what an Artek engineer does with your team, and what you have afterwards.

Guided setup →

Trust

What it reads, what leaves the gateway, permissions, spend caps and the record of every call.

Trust →

FAQ

Versions, data handling, model costs, local models, licensing and setup.

FAQ →


Requirements and setup

What it runs on

Ignition 8.3.4 or later8.3.8 is the verified platform, with Perspective. Reporting for the PDF report and the Artek MQTT suite for MQTT output are optional.
A databaseSQL Server, MariaDB, MySQL or PostgreSQL, with the Artekis tables applied by your DBA. Anomaly Watch also needs tag history on what it screens.
A modelAn Anthropic or OpenAI key, or an OpenAI-compatible endpoint, hosted or on your own network. Ask needs a model that supports tool calling.

The gateway’s Status page shows eight setup steps judged from the gateway’s real state: database tables, a connection and secret, the license or trial, the module, configuration, the views pack, playbooks and the optional incident trigger. In the four hours included, an Artek engineer installs it with your gateway admin, drafts the model for your first area with your controls engineer, sets the schedules, runs the first reports with you and sits with your supervisors in Ask. Most teams use them across two or three remote sessions in the first two weeks.

What it costs to run

Model calls measured in cents

You pay your model provider directly, with daily caps per provider, per run and per user. On our acceptance run, against a 125,000-event alarm journal and a 19,900-tag provider:

Under $0.50for a 30-day Alarm Health run, in about a minute
About two centsfor a 24-hour Gateway Health run
$0.03 to $0.12for a question in Ask, answered in 12 to 17 seconds

A gateway producing 30 to 50 alarm candidates a day runs about $15 to $24 a month in model calls; a small gateway a few dollars. A model on your own network costs nothing per call.

Next step

Tell us about your gateway and what you want to find out.

A member of our team reads every message and replies, usually the same day.

Contact us →