Widgelix Actions watch every reading from your devices and respond on their own: notify the right people, call your systems, or send a command back to the device. Rules fire on sustained situations, not one noisy value, so your team trusts every alert. (Actions were formerly called the Rule Engine.)
Most IoT platforms stop at “if value > X, send email”. In the field that means a flood of alerts from one spiky reading, a value that hovers around the limit, or a gateway outage that makes every sensor look offline. Teams start ignoring notifications, and the one alert that matters is missed.
Pick an asset template and a data field (temperature, CO₂, level, battery), select assets from a list or on a map, or choose All to cover the whole fleet. You can also watch connectivity: asset offline, or gateway offline / back online.
Compare with a constant, a range, the live value of another device, or a geofence. Add a reset threshold and a persist duration so the action reflects a real, sustained condition.
Send an email, SMS, Slack, Discord or SIGNL4 alert, call any webhook, or send a downlink command to a valve, relay or thermostat. Several responses can run together.
Every control below exists to cut false alarms and make automation safe to run without a human in the loop.
Up to 10 statements per action, joined with AND or OR. Combine readings from different devices in one action.
Fire only when a condition holds for a set time, for example 10 minutes above 30 °C. Ignore short spikes without resetting the timer, and reset after a prolonged recovery.
Trigger above 30 °C, re-arm only below 27 °C. Values that hover near the limit no longer flap between alarm and normal.
Make “sensor offline” depend on “gateway offline”. While the parent is active, child alerts stay silent, so you get one alert about the root cause instead of hundreds. A hold window handles near-simultaneous triggers.
Limit an action to business hours or overnight windows on a weekly timeline, cap execution to once per minute, hour or day, or fire once until the condition clears.
Draw circles or polygons on a map and fire when a tracked asset enters or leaves an area. Combine with temporal control to ignore a single stray GPS fix.
Insert live values and asset tags (floor, room, address) into messages. One action for a whole template still tells the recipient exactly which device, where, and what value.
An action can send a downlink back to a device: open an irrigation valve, close a water shut-off, switch a relay. Sensing and actuation run on the same platform.
Statements show the device, the comparison and the persist duration on one card, so anyone on the team can review what an action does before it goes live.
| Condition types | Compare with constant, compare with another asset, geofence, asset offline, gateway offline / online |
|---|---|
| Operators | <, ≤, =, ≥, >, ≠, in range, out of range |
| Conditions per action | Up to 10, joined with AND / OR |
| False-alarm controls | Persist duration, ignore short spikes, reset after prolonged violation, switching threshold range |
| Frequency controls | Trigger once; execute every 1 min, 5 min, 15 min, 30 min, 1 h or 1 day |
| Scheduling | Weekly timeline in 30-minute steps, business-hours preset |
| Responses | Email, SMS, Slack, Discord, SIGNL4, Webhook, Downlink, Event only |
| Severity | Info, Warning, Critical |
| Escalation | Optional escalation chain with acknowledgement deadlines |
“If soil moisture stays below 30 % for 2 hours during daytime, open the irrigation valve with a downlink and notify the field team on Slack.”
No. Actions are configured in a visual editor: pick the devices, the condition and the response. Code is only needed if you want custom payload decoding, which lives in the asset template.
Use temporal control so a condition must hold for a set time, ignore short spikes, and add a switching threshold range so the action only re-arms after the value has clearly recovered. Rate limits and trigger-once further reduce repeat notifications.
Yes. Select All assets of a template and use tag placeholders such as floor or room in the message. Each alert still names the exact device and location.
Yes. The Downlink action sends a predefined command to a device, for example to close a valve or switch a relay, on integrations that support downlinks.
Email, SMS, Slack, Discord, SIGNL4 and generic webhooks, plus in-app events. Webhooks let you connect ticketing systems, BMS or any HTTP API.
Yes. Through the Widgelix MCP server, agents such as Claude or ChatGPT can create, clone, enable and check actions with the permissions of the connected user.
Tell us what you monitor and what should happen when something goes wrong. We'll show you the actions that fit your deployment.