I wanted an app that shows me exactly what I need while driving. Not Google Maps for pedestrians, not another fitness tracker. A real HUD – directly on the phone, logically arranged, without distraction.<\/p>\n
The starting point was a specific situation: bad weather, backlight, unfamiliar route. Traffic lights recognized too late, speed limit missed, no good overview. The smartphone lies on the dashboard and shows 'nothing helpful.'<\/p>\n
DriverHUD is designed to change that. Real-time speedometer, current speed limit from offline maps, early warning for danger zones \u2013 and for drivers with color vision deficiency, an AI that recognizes traffic lights and displays them clearly without needing to focus on the color.<\/p>\n
There are two tiers. Light runs on everything from Android 8: speedometer, GPS quality, speed limits for Hesse from OSM data, danger zone warnings. This is sufficient for everyday use and is free of charge.<\/p>\n
Pro adds live camera analysis on top. A YOLO model recognizes traffic lights in real-time \u2013 color, state, confidence. This requires modern hardware (S22 class or comparable) because computer vision is CPU-bound and the SoC determines latency. Pro is planned to start at \u20ac3.99 per month, running via Play Billing once the store rollout occurs.<\/p>\n
Android because the development path is more direct \u2013 no review gate for early testing, ADB sideloading works immediately. iOS will not arrive before the first stable Android release. Kotlin and Compose because the stack is modern and Claude Code can work well with it \u2013 the app is created in Claude Code sessions, not in months-long sprints.<\/p>\n
Prototype and initial tests are coming tomorrow.<\/p>","protected":false},"excerpt":{"rendered":"