Receipt scanner for Spanish supermarkets that respects your privacy.
Scan receipts from Mercadona, Carrefour, Lidl, and more. Track spending. Keep your data local.
- 100% on-device: ML Kit OCR runs locally. No cloud uploads.
- Any format: Document scanner, gallery, or PDF.
- Smart categories: Auto-sorts items. Learns from your corrections.
- Spanish supermarkets: Pre-trained templates for Mercadona, Carrefour, Lidl, and more.
- Multi-language: English & Spanish UI.
Capture: Document scanner (auto-crop & enhance), gallery import, PDF parsing via ML Kit OCR.
Smart Parsing: Chain-specific templates detect Mercadona, Carrefour, Lidl, Eroski, Dia, Consum, Alcampo, Aldi, HiperDino by NIF/name patterns.
Organize: 11 built-in categories, auto-categorization, user learning, store detection.
Analyze: Monthly trends, category breakdowns, store comparisons, top items.
Privacy: Offline-first, local SQLite, no tracking, no ads. Full policy: jbgsoft.com/cestapp/privacy.
Details: Dark mode, Spanish regional formats (EUR, DD/MM/YYYY, decimal comma), animations that respect Reduce Motion.
git clone https://github.com/jaimebg/CestApp.git
cd CestApp
npm install
# Dev build required (ML Kit needs native code)
npx expo run:ios # or
npx expo run:androidExpo Go won't work. You need a development build.
| Category | Technology |
|---|---|
| Framework | React Native 0.86.2 + Expo SDK 57 |
| Navigation | Expo Router |
| Styling | NativeWind v4 |
| Database | Drizzle ORM + expo-sqlite |
| State | Zustand v5 |
| OCR | ML Kit (@infinitered) |
| Custom parser + pako | |
| i18n | i18next |
| Animations | Reanimated v4 |
| Charts | react-native-gifted-charts |
app/ # Screens (Expo Router)
src/
components/ # UI components
db/ # Schema & queries
services/ # OCR, PDF, capture, storage
store/ # Zustand state
i18n/ # Translations
- Detect Chain: Identifies supermarket by NIF (tax ID), store name patterns, or fingerprints (brand names like "Hacendado" for Mercadona)
- Apply Template: Uses chain-specific parsing rules (layout, item patterns, OCR corrections)
- Extract: ML Kit OCR for images, custom parser for PDFs
- Categorize: User corrections → keyword matching → default
Learning is store-aware: same item can have different categories at different stores.
Supported Chains: Mercadona, Carrefour, Lidl, Eroski, Dia, Consum, Alcampo, Aldi, HiperDino (with IGIC tax for Canarias).
- Fork → branch → commit → PR
- TypeScript strict mode
- NativeWind for styling
useTranslation()for all text- Test iOS + Android + dark mode
MIT. See LICENSE.
Made with care by JBGSoft - Jaime Barreto 🧡