I used AI, especially vibe coding, throughout MetroPing's development. It helped me move much faster as an independent developer, but speed and reliability turned out to be very different problems.
What AI genuinely helped with
AI helped generate ideas, create prototypes, debug problems, and automate repetitive development work. It reduced the distance between an idea and a testable version.
That speed mattered for a side project because it let me try more approaches without a large team.
What AI could not validate
It could not sit inside a moving Delhi Metro train while the phone was locked. It could not reproduce every Android manufacturer's background restrictions or reveal exactly where underground location would weaken.
Those failures appeared only when the app met real devices, real settings, and real journeys.
The lesson
AI can help an independent developer reach the testing stage faster. It does not remove the testing stage.
MetroPing became more reliable through debugging, patience, volunteer feedback, and repeated travel across actual Delhi Metro lines.
Faster code is not automatically a validated product.
AI accelerated development. Reliability still came from using MetroPing in the conditions it was built for.
Try it on a real Delhi Metro journey.
MetroPing is not affiliated with DMRC. Use it as an additional reminder and share honest feedback about the experience.
Try MetroPing