Builder made OurCar, a Flutter + Pocketbase app to solve car-sharing logistics for his family, covering gas tracking, location, and scheduling.
Key Takeaways
Stack: Flutter with Riverpod and Auto Route on frontend, Pocketbase backend with hooks for push notifications; chosen for prior familiarity and proved sufficient.
Gas tracking works by combining trip odometer and trip efficiency readings to calculate fuel consumed per trip, entered manually at car return.
GPS tracking was deliberately excluded: live location drains battery and exposes private trip destinations to all family members; only parked location is shared.
Flutter’s lack of automatic platform-adaptive widgets forced reliance on the flutter_platform_widgets package, resulting in spaghetti layout code across Material and Cupertino.
Distribution via App Store and Play Store beta channels was painful upfront, especially Apple notarization; author eventually paired wife’s phone directly to developer account.
Hacker News Comment Review
Commenters split on manual data entry: skeptics argue it is too unreliable at scale and that OBD-II integration would be necessary for real-world use; author counters the app targets trusted groups where friction reduction matters more than automation.
Flutter cross-platform native feel drew scrutiny; one reply argued Expo with expo-ui now outperforms Flutter for matching native iOS/Android look and feel, especially post-Liquid Glass.
Broader thread surfaces a recurring pattern: small-audience family tools are low-stakes enough to absorb LLM-generated code for uninteresting flows while keeping manual effort for the parts worth learning.
Notable Comments
@tantalor: “Good example of over-engineering. This could be easily solved with a pencil and notepad.”
@fmajid: argues manual entry fails without OBD-II fuel data plus presence tracking for reliability at any real scale.
