The WMS
A stalled dock inside a booked slot becomes a flag against that slot, where the people who plan the yard will actually see it.
Channel
Your systems should hear it
before a person does.
Every livee alert can POST a structured payload to an endpoint you own. It is the channel that turns a camera event into something your WMS, your incident tooling or your own middleware can act on — without anybody retyping what the camera saw.
The payload
What arrives at your endpoint.
Stable field names, an event id you can deduplicate on, and the measurement that justified the alert.
event_id— unique and stable, so a retry is not a second incidentoccurred_at— ISO 8601, UTC, to the secondsiteandsource— which hall, which camera, by the labels you gave themalert— the rule that fired, with its type: motion, stall or proximityobject— class and confidence, where the alert type has onemeasure— distance against threshold, or the stall duration against the graceclip_url— the seconds around the event, for the person who has to review itchannels— everything else this event also notified
Integrating
Three steps, and it is in your stack.
- 01
Give us an endpoint
Any HTTPS URL you control — your WMS, an incident tool, a chat relay, a function you wrote this morning. One per alert, or one for the whole site.
- 02
Verify the signature
Deliveries are signed, so your endpoint can prove a POST came from livee and not from someone who found the URL.
- 03
Answer 2xx
A non-2xx or a timeout is retried and logged. You can see what was delivered, when, and what your endpoint said about it.
Patterns
What customers wire it to.
Incident tooling
A keep-out breach opens a ticket with the clip attached, so the safety review has the footage rather than a recollection.
A chat relay
Your own middleware forwards the event into the channel the shift already lives in. livee does not need to know which one that is.
Your own analytics
Events as data: how often a crossing rule fires, at which hour, on which camera. The pattern is usually more useful than any single alert.
Signed, retried, logged.
Request a demoFAQ
About the API channel.
Do I need an integration project?
For most sites, no. If you can receive an HTTPS POST and read JSON, you can consume livee events. The work is usually deciding what your system should do with them, not wiring them up.
Can different alerts POST to different endpoints?
Yes. Channels are attached per alert, so a stall can go to the WMS while a keep-out goes to your incident tool.
What about outbound restrictions?
If your network only accepts traffic from inside itself, that points at an on-site deployment, or at the Ethernet channel for the machine-facing part. It is one of the first things we settle on the call.