How it works
From engine signal to dispatch alert in real time
The install
One device per truck, connected to the CAN bus at the diagnostic port. A fleet of fifty is typically live in a day, and trucks keep working while it happens.
The data
The CAN bus already carries everything worth knowing, coolant temp, oil pressure, fault codes, fuel rate, braking, idling, speed. The Canary reads all of it continuously and streams it to the platform, giving each truck a live health record.
The prediction
Machine learning models compare every truck against its own history and against the whole fleet. When a pattern starts matching the run-up to a known failure, the system flags it, scores the risk, and estimates the service window you have left.
The alert
Dispatch sees it first. Each alert names the truck, the component, the severity, and the recommended action, so the decision takes a phone call instead of a diagnosis. Alerts route by email, SMS, and the dashboard.
The conversation
Drivers don't file tickets, they talk. A driver mentions a warning light or a sluggish grade climb, hands-free and DOT-compliant, in any of 12-plus languages, with zero training. The system weighs what the driver said against live vehicle signals, the route, the weather, and remaining drive time, and hands dispatch a diagnosis instead of a hunch.
The follow-through
A warning that just creates work isn't a favor. When a fault is confirmed, the platform can schedule the service stop along the truck's active route, get the parts on order, and line up a backup vehicle before the driver ever pulls in. Dispatch approves the plan; the system does the legwork.
The payoff compounds
Every mile makes the models sharper. Every fault confirmed makes the next warning earlier. Fleets that run the Canary longest get the widest service windows.
See the whole loop on your own fleet's terms
Twenty minutes, live dashboard, real alerts.