Skip to main content

How Haltless Works

How Haltless works​

Haltless sits between your machines and your team. Sensor data flows in, anomaly detection watches it in real time, and the signals come back out as alerts you can act on: in the dashboard, over the API, on a live WebSocket stream, or delivered straight to your own systems.

This page is a conceptual map. You don't operate any of this infrastructure yourself; you integrate with the hosted platform at https://api.haltless.io/api/v1.

The path your data takes​

Machines and sensors on your plant are read by Haltless Edge on your network, or sent by your own software to the direct ingestion API, over outbound HTTPS only. In the Haltless cloud, readings are ingested, checked by anomaly detection, and raise alerts, which reach you through the dashboard, the REST API, a WebSocket stream and webhooks.Machines and sensors on your plant are read by Haltless Edge on your network, or sent by your own software to the direct ingestion API, over outbound HTTPS only. In the Haltless cloud, readings are ingested, checked by anomaly detection, and raise alerts, which reach you through the dashboard, the REST API, a WebSocket stream and webhooks.

The pieces​

Your machines and sensors​

Everything starts with the equipment on your floor: temperature probes, vibration sensors, pressure gauges, and whatever else reports the health of a machine. Haltless calls the connection that data comes from a System: a PLC, an OPC-UA server, an MQTT broker, or a folder of exported files. Each System has an identifier you choose and owns every tag it sends. A machine lives inside a System and is made of the tags assigned to it, so one PLC that runs a whole line can feed several machines. Each reading carries a metric name, a value, a unit, and a timestamp. See Systems.

Getting data in​

There are two ways to move readings from your equipment into Haltless:

  • The Edge Agent. A lightweight agent you run near your machines. It connects to your sources (PLCs over OPC-UA, Modbus, Siemens S7, EtherNet/IP, Mitsubishi MELSEC, Omron FINS, or Beckhoff ADS, BACnet/IP building controllers, MQTT or Sparkplug B brokers, an AVEVA PI historian, or CSV and JSON files), collects readings, and streams them to Haltless. This is the usual choice for continuous production data. See the Edge Agent overview.
  • Direct ingestion. Send readings yourself with POST /api/v1/ingest. This is ideal when your data already lives in a system that can make HTTP calls, or when you want to try things quickly. See the direct ingestion guide.

Both paths carry the same kind of readings and feed the same pipeline.

Anomaly detection​

As readings arrive, Haltless compares each one against what "normal" looks like for that machine and metric. When a value stays outside its expected range for a few readings in a row, Haltless raises an alert. You can shape when and how alerts fire with alert rules, and Haltless also surfaces forward-looking predictions about equipment health.

Getting signals out​

Once an alert exists, you can consume it in whichever way fits your workflow:

ChannelBest forLearn more
DashboardPeople watching the fleet in a browser
REST APIPulling alerts and data on your scheduleAlerts
WebSocket streamReacting the instant something changesWebSocket integration
WebhooksPushing alerts into your own systemsWebhook setup

Webhook deliveries carry the custom headers you configure on the channel, such as a shared secret, so your receiver can verify they genuinely came from Haltless.

From alert to action​

Alerts are the starting point, not the end. Around them, Haltless gives you the maintenance building blocks to close the loop: work orders, maintenance records, downtime logging, OEE, and more, all reachable through the API.

Security and access​

  • Encryption everywhere. Traffic to the API is protected with TLS in transit, and sensitive data is encrypted at rest.
  • Two kinds of credentials. API keys authenticate data coming in from your machines; Bearer access tokens authenticate people and applications reading and managing data. See Authentication.
  • Role-based access. Your role controls what you can see and do; account admins manage your team's roles.

Next steps​