Native mobile · Applied AI
Leitor Hidrômetro
A field workflow for capturing water meters, extracting readings with AI and organizing reviewed results for operations teams.
The problem
Manual meter collection is repetitive, error-prone and often performed with intermittent connectivity. The product needed to help field operators capture legible evidence, continue working offline and review uncertain readings before they entered operational records.
The system
The Android app uses Kotlin, Jetpack Compose and CameraX for the primary field workflow, with a native Swift implementation for iOS. A FastAPI service supports AI-assisted extraction using vision-language or OCR pipelines, while a local queue protects in-progress work when connectivity drops.
Product delivery
The experience extends beyond recognition: it organizes customers and meters, supports human confirmation, produces spreadsheet-friendly outputs and connects commercial flows such as wallet and Pix support. Leitor Hidrômetro is available on the App Store and Google Play.
Result
A capture-to-review pipeline combining camera guidance, offline queues, AI extraction and exportable operational records.
Project gallery
