VemuniApp

Vemuni

Vemuni is a SwiftUI iPhone app that makes restaurant menus easier to read and checks restaurant bills for fees and possible double tipping.

What is included

• Menu and receipt photo capture or photo-library selection
• Automatic language detection and English translation
• Plain-language dish explanations, prices, dietary indicators, and cautious allergen hints
• Receipt subtotal, tax, service-charge, gratuity, and total extraction
• Tip calculations based on the pretax subtotal when visible
• Google, Apple, and email/password authentication through Firebase
• Firebase Analytics events for acquisition and funnel measurement without menu, receipt, image, or email contents
• Ten introductory scans followed by one free scan per UTC day, enforced server-side
• Firebase HTTPS function that keeps the OpenAI API key off the phone
• Per-scan AI usage records and lifetime token totals in Firestore, without menu text or images
• Bundle ID bestvaluellc.apps.vemuni

Open the iPhone app

Open Vemuni.xcodeproj in Xcode. Select your Apple development team for the Vemuni target, confirm the Sign in with Apple capability, allow Swift Package Manager to resolve Firebase and Google Sign-In, and run on an iPhone or simulator.

The included GoogleService-Info.plist is configured for Firebase project vemuni-d4f87. The app’s Info.plist already contains the matching Google callback URL and deployed-function URL.

Deploy the scan function

From this folder:

cd functions
npm install
cd ..
firebase functions:secrets:set OPENAI_API_KEY
firebase deploy --only functions,firestore:rules

The endpoint is analyzeDiningPhoto in us-central1. It verifies a Firebase ID token, sends the image to the OpenAI Responses API using structured output, records scan allowance counters and token-usage metadata in Firestore, and does not save the uploaded image. Usage records contain the response ID, scan type, page count, model, input, cached-input, output, reasoning, and total token counts. They do not contain menu text, receipt values, or images.

Analytics events

Vemuni records scan_opened, scan_photo_selected, scan_started, scan_completed, scan_failed, menu_item_opened, auth_attempt, and Firebase’s standard login event. Parameters are limited to scan type, photo source, result counts, detected-language label, broad error category, and boolean feature flags. Never add menu text, receipt values, email addresses, image data, or other user-provided content to Analytics.

Before App Store submission

• Add Terms of Use and Privacy Policy destinations.
• Add in-app account deletion because Vemuni supports account creation.
• Enable and enforce Firebase App Check to reduce unauthorized API use.
• Replace per-install local quota messaging with the server-returned quota state.
• Test Google and Apple authentication on a signed physical device.
• Add accessibility and localization QA.

Menu explanations and allergen indicators are informational. Users should always verify ingredients and allergens with the restaurant.