Dual-Ecosystem Product: Native Android Production App + Cross-Platform Progressive Web App (PWA) + Firebase Cloud Sync & Auth + Production ASO & SEO Pack.
DocuSnap AI is an intelligent document vault engineered to solve the everyday chaos of lost medical prescriptions, pharmacy slips, electricity/gas utility bills, and grocery receipts.
Rather than forcing users to manually open a scanner app, DocuSnap AI introduces Zero-Click Media Ingestion:
ContentObserver monitors gallery additions (MediaStore.Images.Media.EXTERNAL_CONTENT_URI).FileProvider with automatic high-res capture and OCR. ┌──────────────────────────────────────────────┐
│ Android Gallery Photo / Snap │
└──────────────────────┬───────────────────────┘
│
▼
┌─────────────────────────────┐
│ MediaObserverService │
│ (ContentObserver) │
└──────────────┬──────────────┘
│
▼
┌─────────────────────────────┐
│ MediaIngestionWorker │
│ (WorkManager Coroutine) │
└───────┬──────────────┬──────┘
│ │
┌───────────────────────────┘ └───────────────────────────┐
▼ ▼
┌───────────────────────┐ ┌──────────────────────┐
│ Google ML Kit OCR │ │ Resolution & Hash │
│ (100% On-Device) │ │ SHA-256 Dedupe │
└──────────┬────────────┘ └──────────────────────┘
│
▼
┌───────────────────────┐
│ Gemini Flash Reasoning│ ───► Structured Entities: Doctor, Medicines, PKR Amounts, Vendor
└──────────┬────────────┘
│
▼
┌───────────────────────┐ ┌──────────────────────┐
│ Room DB (SQLite FTS5)│ ◄─────────────────────────────────────────── │ Conversational UI │
│ Offline Search Index │ │ "Ask Anything" Box │
└──────────┬────────────┘ └──────────────────────┘
│
▼
┌───────────────────────┐
│ Firebase Cloud Sync │ ◄───► Real-Time Firestore & Storage
└──────────┬────────────┘
│
▼
┌────────────────────────────────────────────────────────┐
│ Cross-Platform PWA (iOS Safari / Desktop / Mac / Win) │
│ - HTML5 Canvas Downscaling │
│ - Native iOS "Add to Home Screen" Bottom Sheet Modal │
│ - Zero-Cost Static Hosting on GitHub Pages │
└────────────────────────────────────────────────────────┘
DocuSnapAI/
├── android/ # Native Android Production Application
│ ├── build.gradle.kts # Project build script
│ ├── settings.gradle.kts # Gradle settings
│ ├── gradle.properties # JVM & AndroidX properties
│ └── app/
│ ├── build.gradle.kts # Compose, Room FTS5, ML Kit, Gemini, Firebase BOM
│ ├── proguard-rules.pro # ProGuard rules for Room, ML Kit & Gemini
│ └── src/main/
│ ├── AndroidManifest.xml # Permissions (READ_MEDIA_IMAGES, FOREGROUND_SERVICE)
│ ├── java/com/docusnap/ai/
│ │ ├── DocuSnapApplication.kt
│ │ ├── ai/
│ │ │ └── GeminiReasoningEngine.kt # Structured JSON & Conversational Search
│ │ ├── data/
│ │ │ ├── local/
│ │ │ │ ├── AppDatabase.kt # Room Database with FTS virtual table
│ │ │ │ └── DocumentDao.kt # FTS5 MATCH queries & CRUD
│ │ │ └── model/
│ │ │ ├── DocumentEntity.kt # Primary entity schema
│ │ │ ├── DocumentFtsEntity.kt # Room FTS4/5 entity
│ │ │ └── DocumentDto.kt # DTOs & JSON serialization
│ │ ├── ocr/
│ │ │ └── OcrEngine.kt # Bundled Google ML Kit Text Recognition
│ │ ├── service/
│ │ │ ├── BootReceiver.kt # Restarts observer service on device boot
│ │ │ ├── MediaIngestionWorker.kt # CoroutineWorker with aspect ratio filter
│ │ │ └── MediaObserverService.kt # Foreground Service + ContentObserver
│ │ ├── sync/
│ │ │ └── FirebaseSyncRepository.kt # Auth, Firestore & Storage sync
│ │ └── ui/
│ │ ├── MainActivity.kt # Compose activity & permissions launcher
│ │ ├── DocuSnapNavHost.kt # Navigation graph & bottom bar
│ │ ├── DocuSnapViewModel.kt # StateFlow & conversational orchestrator
│ │ ├── screens/
│ │ │ ├── ConversationalSearchScreen.kt # "Ask Anything" (Roman Urdu/English)
│ │ │ ├── DocumentInspectorScreen.kt # Deep-zoomable viewer & copyable tags
│ │ │ ├── SettingsScreen.kt # On-Device vs Gemini & battery settings
│ │ │ └── SmartVaultScreen.kt # Categorized grid of documents
│ │ └── theme/
│ │ ├── Color.kt
│ │ ├── Theme.kt
│ │ └── Type.kt
│ └── res/ # Drawables, colors, strings, extraction rules
│
├── web/ # Cross-Platform Progressive Web App (PWA)
│ ├── package.json # React 18, Vite, Tailwind CSS, Lucide, Firebase JS
│ ├── vite.config.ts # Vite + VitePWA configuration
│ ├── tailwind.config.js # Theme colors matching Android Material 3
│ ├── tsconfig.json # TypeScript bundler configuration
│ ├── index.html # PWA meta tags & iOS Safari viewport fit
│ ├── public/
│ │ ├── 404.html # GitHub Pages SPA routing fallback script
│ │ ├── manifest.json # Full PWA Web App Manifest
│ │ └── favicon.svg # DocuSnap vector brand asset
│ └── src/
│ ├── App.tsx # Main container, tabs, and real-time synchronization
│ ├── firebase.ts # Firebase Auth, Storage, & Firestore snapshot listeners
│ ├── gemini.ts # Gemini Flash multimodal vision & Roman Urdu reasoning
│ ├── types.ts # TypeScript models matching Room Entity
│ └── components/
│ ├── ConversationalSearch.tsx # "Ask Anything" search with animated AI answer
│ ├── SmartVault.tsx # Categorized grid with filter chips & amounts
│ ├── DocumentInspectorModal.tsx # Deep-zoom image inspector with one-click copy
│ ├── UploadModal.tsx # Client-side HTML5 Canvas downscaler & camera intake
│ ├── IosInstallPrompt.tsx # Native iOS Safari "Add to Home Screen" bottom sheet
│ └── Navbar.tsx # Brand header & navigation controls
│
├── firebase/ # Cloud Backend Contracts & Rules
│ ├── firebase.json # Hosting & Firestore configuration
│ ├── firestore.rules # Security rules: users/{userId}/documents/{docId}
│ ├── firestore.indexes.json # Composite query indexes
│ └── storage.rules # Rules for /images/ and /thumbs/
│
├── play_store_assets/ # Google Play Store Production ASO
│ ├── title.txt # Title (29/30 characters)
│ ├── short_description.txt # Short Description (79/80 characters)
│ ├── full_description.txt # Keyword-rich 4000-character description
│ ├── listing_metadata.json # Categories, ratings, contact info
│ └── release_notes.txt # Release notes for v1.0.0
│
├── web_seo/ # Web SEO & Search Engine Optimization Pack
│ ├── index.html # Complete OpenGraph, Twitter Card & JSON-LD
│ ├── robots.txt # Search engine crawler directives
│ └── sitemap.xml # XML Sitemap with daily change frequency
│
└── README.md # Comprehensive product documentation
In android/.../service/MediaObserverService.kt:
ContentObserver on MediaStore.Images.Media.EXTERNAL_CONTENT_URI.MediaIngestionWorker via ExistingWorkPolicy.APPEND_OR_REPLACE.MediaIngestionWorker decodes image bounds first without loading full bitmaps into RAM:
width < 380 or height < 380, discards (small stickers, low-res thumbnails).aspectRatio > 4.5, discards (thin banners, dividers).When Gemini Deep Reasoning is active, extracted text is analyzed using the following structured schema:
{
"document_type": "Medical Prescription | Expense Receipt | Utility Bill | Bank Slip | General Document",
"title": "Short descriptive title (e.g., Blood Test Report - Chughtai Lab)",
"summary": "1-2 sentence conversational summary",
"key_entities": {
"doctor_name": "string or null",
"medicines": ["string"],
"total_amount": 1250.00,
"vendor_or_hospital": "string or null",
"date": "YYYY-MM-DD"
},
"searchable_tags": ["tag1", "tag2"]
}
documents_fts MATCH :query).display: "standalone", viewport-fit=cover, custom dark theme colors).IosInstallPrompt: step-by-step bottom sheet guiding the user to tap Share -> Add to Home Screen.android/ directory in Android Studio.GEMINI_API_KEY in android/app/build.gradle.kts or local.properties.google-services.json from the Firebase console in android/app/../gradlew assembleDebug
web/ directory:
cd web
npm install
npm run build
web/dist/.# Push dist directory to gh-pages branch
git subtree push --prefix web/dist origin gh-pages
Note: public/404.html ensures all deep SPA routes redirect seamlessly without 404 errors on GitHub Pages.
From the project root:
firebase deploy --only firestore:rules,storage:rules
DocuSnap AI is released under the MIT License.