Build shape
Released consumer expense app
AI expense tracking app
Receipt Snap
An AI expense app that uses Gemini to read receipts and carries the flow through review, save, and stats.
FinanceAIOCR
Product brief
The bigger problem was not OCR itself but giving users a comfortable expense-review flow before save.
Receipt Snap is an expense app built around capture, crop, Gemini extraction, and confirm-before-save. The product is intentionally review-first rather than fully automatic.
The repo shows camera and gallery input, a confirmation screen, history, stats, CSV export, reminders, offline storage, and Google-backed sync. It is a grounded consumer AI product focused on repeat use.
Scope
Problem
Problem
For receipt apps, review-before-save matters more than raw recognition. The product needed user correction, not blind automation.
Implementation
I combined camera and gallery input, cropping, Gemini extraction, confirmation, history, stats, CSV export, and reminders in a Flutter app with local storage.
Best fit
This fits consumer apps where AI capture helps, but the final review still belongs to the user.
Technical structure
Product evidence
Primary stack
Flutter · Riverpod · GoRouter
Data & sync
Drift SQLite · Firebase Auth · Firestore
Operational surface
AdMob · In-app purchase · reminders · CSV export
Stack
App client
Flutterhooks_riverpodflutter_hooksgo_routerget_it
Data and AI
DriftFirebase AuthCloud FirestoreCloud FunctionsFirebase AIGoogle ML Kitimage_cropper
Operations and monetization
FCMflutter_local_notificationsGoogle Sign-InAdMobin_app_purchasefl_chart
Integrations
Firebase AuthCloud FirestoreCloud FunctionsFirebase AI (Gemini)ML Kit Document ScannerGoogle Sign-InFCMAdMobGoogle Play Billing
Screens
Product evidence
App
Need a similar product, internal tool, or launch-ready first version?
Send the current problem, scope, or a reference link. The next step can start from product structure and release scope.