SnapOCR

A mobile application that turns photos of documents into searchable, editable PDF files using Mistral AI's OCR engine. Point your camera at any document -- printed text, handwritten notes, receipts, forms -- and get a clean PDF with selectable text in seconds.

Architecture

frontend/          React Native (Expo) mobile app
backend/           Node.js / Express REST API
docs/              Architecture and context documentation

The frontend captures or selects images and sends them to the backend, which delegates OCR processing to Mistral AI's vision models. The extracted text is post-processed for accuracy (symbol correction, layout preservation) and rendered into a structured PDF using PDFKit.

Tech Stack

Layer	Technologies
Mobile	React Native 0.76, Expo, TypeScript, Expo Router, React Native Paper
Backend	Node.js 18, Express, MongoDB (Mongoose), JWT auth
OCR	Mistral AI OCR API (mistral-ocr-latest, mistral-large-latest)
PDF	PDFKit with layout-aware rendering (headings, lists, tables, columns)
Auth	JWT + bcrypt, Apple Sign-In (expo-apple-authentication)
Payments	RevenueCat (react-native-purchases) for iOS/Android subscriptions
Encryption	Client-side XOR file encryption before upload
Deploy	Railway (Nixpacks)

Features

• Camera and gallery upload -- scan documents directly or import from photo library
• Intelligent OCR -- recognizes printed text, handwriting, tables, formulas, and special characters
• PDF generation -- produces structured PDFs preserving headings, lists, and layout
• Post-processing -- fixes common OCR errors (math symbols, bullet points, image artifacts)
• Handwriting mode -- specialized processing pipeline tuned for handwritten documents
• Client-side encryption -- files are encrypted before upload for privacy
• Subscription system -- free tier (3 docs/day), premium, family, and business plans
• Document history -- browse, re-download, and share previously processed documents
• Multi-language -- i18n support with language switching
• Dark mode -- full dark/light theme support
• Apple Sign-In -- native authentication on iOS
• Family sharing -- manage a shared subscription with up to 4 members

Getting Started

Prerequisites

• Node.js 18+
• MongoDB (Atlas or local)
• Mistral AI API key (platform.mistral.ai)
• Expo CLI (npm install -g expo-cli)

Backend

cd backend
cp .env.example .env    # fill in your credentials
npm install
npm run dev

Frontend

cd frontend
npm install
npx expo start

Set EXPO_PUBLIC_API_URL in your environment to point to your backend (defaults to http://localhost:3001/api).

Project Structure

SnapOCR/
├── backend/
│   ├── config/             Database and upload configuration
│   ├── controllers/        Route handlers (auth, documents, subscriptions)
│   ├── middleware/          Auth, validation, error handling
│   ├── models/             Mongoose schemas (User, Document, ProcessingJob, FamilyGroup)
│   ├── routes/             Express route definitions
│   ├── services/
│   │   ├── ocrService.js           Core OCR + PDF pipeline
│   │   ├── pdfService.js           Image-to-PDF conversion
│   │   ├── processingService.js    Job orchestration
│   │   ├── subscriptionService.js  Plan management and billing
│   │   ├── translationService.js   OCR with translation
│   │   ├── tableExtractionService.js
│   │   ├── classificationService.js
│   │   ├── qrBarcodeService.js
│   │   └── imageEnhancementService.js
│   ├── utils/              Token generation, validators
│   └── server.js           Entry point
├── frontend/
│   └── app/
│       ├── (auth)/         Sign-in, sign-up screens
│       ├── (app)/          Main app screens (dashboard, upload, history, profile, etc.)
│       ├── components/     Shared UI components
│       ├── constants/      API config, colors, routes, theme
│       ├── contexts/       Auth, Document, Subscription, Language, DarkMode providers
│       ├── hooks/          useAuth, useDocuments
│       ├── services/       API client and service layers
│       ├── types/          TypeScript type definitions
│       └── utils/          i18n, encryption, theming
└── docs/                   Architecture documentation

API Endpoints

Auth

• POST /api/auth/register -- create account
• POST /api/auth/login -- email/password login
• POST /api/auth/apple-signin -- Apple Sign-In
• GET /api/auth/profile -- get profile (protected)
• PUT /api/auth/profile -- update profile / change password (protected)

Documents

• POST /api/documents/upload -- upload image for OCR processing (protected)
• GET /api/documents -- list user documents (protected)
• GET /api/documents/:id -- get document by ID (protected)
• GET /api/documents/:id/status -- check processing status (protected)
• DELETE /api/documents/:id -- delete document (protected)

Subscriptions

• GET /api/subscription -- get subscription details (protected)
• POST /api/subscription/trial -- start free trial (protected)
• POST /api/subscription/subscribe -- subscribe to a plan (protected)
• POST /api/subscription/verify-purchase -- verify App Store / Play Store receipt (protected)
• POST /api/subscription/restore -- restore purchases (protected)
• POST /api/subscription/cancel -- cancel subscription (protected)

License

Private project.