Stop 1 · Node map
Every capture is a point on the map.
Each node is one POV capture recorded on consumer smart glasses and positioned by DataRail's own capture app. The map is built from what the capture itself recorded — not from a pin dropped by hand afterwards.
Sequence
Nodes in capture order.
Captured nodes open their reel. Reserved nodes hold their place until a capture fills them.
How a node gets made
Capture in, node out.
Three of these four links work today. The fourth is the one being built, and it is why this page is a working surface rather than a mockup.
Capture
POV video recorded on consumer smart glasses while moving through the city on foot and by rail.
WorkingPosition
DataRail's capture app ties each recording to where and when it was taken, so a reel arrives knowing its own place in the city.
WorkingIngest
Turning an uploaded capture into a node without a human in the loop. Not built yet — nodes are entered by hand for now.
Open edgeNode
The node is published in a plain, readable shape and this page renders it: position, route, reel and all.
WorkingThe capture app
Hand-built, in Java and Kotlin.
The glasses record. Everything that makes a recording into a node comes from DataRail's own capture app — written by hand in Java and Kotlin, no AI-generated code and no framework scaffolding, developed against the Wearables Device Access Toolkit. It is the piece that turns a walk through a city into something a tool can read, and it is the part that is genuinely ours.
No generated code
Written line by line in Java and Kotlin. Nothing scaffolded, nothing autocompleted into place, nothing that nobody can account for.
Capture is the product
The app runs where the recording happens, so a reel is complete the moment it stops — not stitched together from separate sources afterwards.
Plain output
What comes out is deliberately simple to consume, so other tools can build on the same record without needing anything of ours to run.