Most software of this kind does not publish a page like this. It exists because a limit a buyer discovers in week three of a pilot costs more than one they read on the way in, and because every other claim on this site is easier to believe once this page exists.
It is generated from the same registers the product runs on, so it cannot quietly fall behind what is true.
The registers behind this page last changed 2026-09-12. What changed, and when, is on the changelog.
Partial. Unsealed roads are scored on their own terms rather than borrowing a sealed-road method.
The limit: They are handled as one group. We cannot yet report gravel, formed and unformed lengths separately, which some state returns ask for.
Partial. Where a council surveys lanes separately, each lane keeps its own condition and its own recommended treatment.
The limit: We can show a lane needs different work; we cannot yet price it, because nothing records how wide a single lane is.
Experimental. Comparing condition before and after work to estimate what a treatment actually delivered.
The limit: Modelled from a small number of observations rather than measured, off unless a council turns it on, and not a basis for a spending decision yet.
2 commitments are written down as something PaveKeep intends to do and has not yet done. Stating any of them as fact would be a false claim about the business rather than about the product, so they are listed instead.
Generated from the same register, on the same day, in the same shape as this one.