Vehicle
Live today: the root node draws on official Israeli plate lookup, so plate, make, model, year, and registration data populate automatically.
Platform · Foundation
The vehicle health graph
Every car, part, service, and renewal modelled as one connected graph that evolves over time. This is the engineering vision behind the Vehicle Health Score now rolling out in early access, and the substrate future predictions and fleet analytics will read from.
Why a graph
Records are flat. Vehicles are not.
A spreadsheet row can tell you a car exists. It can't tell you that the timing belt is overdue because of this mileage, on this engine, given this service gap. That's why we're building the health graph: vehicles stored as connected entities (components, services, documents, and renewals all linked), so context travels with every data point instead of being lost in separate tables. Here's the design.
- Each vehicle is a node, not a row: connected to its make, model, year, engine, and real registration data.
- Parts and systems hang off the vehicle as their own nodes, each with its own service interval and history.
- Mileage, services, and renewals become time-stamped edges, so the graph remembers how the vehicle got to today.
- Confidence labels are stored alongside facts, so the model never presents an estimate as if it were certain.
Components & systems
In the design: brakes, belts, fluids, filters, and tyres, each modelled with its own wear curve and service interval. Today, intervals live in each vehicle's AI maintenance schedule.
Service history
Live today: every recorded service, repair, and part replacement, time-stamped, with receipt scanning that captures the details for you.
Mileage timeline
Live today: odometer readings over time, turned into a usage pattern that projects when work comes due.
Renewals & documents
Live today: annual test, insurance, and licensing dates tracked per vehicle, so deadlines surface before they lapse.
Model context
Live today: make-, model-, and year-level knowledge (typical intervals and known patterns), applied to every schedule with honest confidence labels.
Where it's headed
A living model, not a snapshot
The goal is a graph that's never built once and filed away. Every new mileage reading, logged service, or renewal will update the connected model, and because relationships are explicit, a single new fact will be able to shift the Health Score and upcoming work across the whole vehicle automatically.
- New readings will re-evaluate affected components in place, with no manual recalculation.
- Changes will propagate along relationships, so one update keeps the whole vehicle consistent.
- History is preserved, so you'll see not just the state today but how it got there.
The substrate everything will read from
The graph is the layer we're building beneath the rest of the platform, so higher-level capabilities can query it instead of re-deriving raw data.
Predictive engine
Today's due-date projection already reads real mileage patterns; the graph will deepen it with component-level intervals.
Recommendation engine
Today's AI schedules will draw on graph state for richer, explainable next steps with honest confidence.
Fleet analytics
On our roadmap: aggregating across many vehicle graphs to surface cost and health trends fleet-wide.
API & integrations
On our roadmap: exposing graph entities and changes so your existing systems stay in sync.
When data is connected instead of filed, context stops getting lost between the columns.
Unbroken Car
Watch the health graph take shape
Get early access and see the foundations working today, or talk to us about helping shape what the graph powers next.