Privacy Policy

This Privacy Policy explains how AVGusto ("we", "us", or "our") collects, uses, and protects your information when you use our mobile application. We are committed to protecting your privacy and ensuring transparency about how we process your data and use device permissions.

Why We Need Camera Access

The camera permission is essential to the core functionality of AVGusto. We use the camera for:

• Meal Photo Capture: Take photos of meals to analyze nutritional content using AI-powered food recognition.
• Real-time Food Analysis: Captured images are processed to identify food items, estimate portion sizes, and calculate nutrition information (calories, protein, carbohydrates, and fats).
• Visual Food Diary: Captured photos are stored locally on your device as part of your personal food diary.

How Camera Access Works

• User-Initiated: The camera is only activated when you explicitly tap the "Scan Meal" button.
• On-Demand: Camera access is temporary and only active during the photo capture process.
• No Background Access: The app never accesses the camera in the background or without your knowledge.

• Saving Meal Photos: After capturing or analyzing a meal, photos are saved to your device's photo library.
• Loading Existing Photos: You can select existing food photos from your gallery to analyze.
• Local Storage: All meal photos and analysis results are stored locally on your device for offline access.

The microphone permission may be requested by the Expo Camera library as a standard permission for camera functionality on Android. AVGusto uses the camera for photo capture only. We do not record audio or video, and the microphone is not actively used.

AVGusto requires an account to use the app. You can sign in via email/password or Google Sign-In through Firebase Authentication. When you sign in, your email address and display name are collected and associated with your account.

Profile Data

You may optionally provide weight, height, age, gender, activity level, and nutrition goals. This data is stored in Firebase Firestore and used solely to calculate personalized calorie and macronutrient targets.

Local-First Architecture

• Meal photos are stored locally on your device and are never uploaded to our servers.
• The app functions offline except for AI analysis, authentication, and sync features.

Cloud Sync

• Your profile, nutrition goals, daily scan counts, scan pack balance, streak data, body-weight entries, and calorie-target calibration data are synced to Firebase Firestore for multi-device access.
• Your food-diary entries (food names, nutrition values, and timestamps — but not the photos) are synced to Firebase Firestore so your history is available across devices and preserved if you reinstall.
• If you set a profile avatar, it is uploaded to Firebase Storage (avatars/{uid}.jpg). Meal photos are NOT uploaded to any server.

AI Processing

• What the AI Does: Meal photos are processed by third-party AI vision models to recognize the foods in the picture and estimate portion size and nutrition. Barcode scans and manual entries do not involve AI.
• Automated, Approximate Output: The resulting calorie and macronutrient values are generated automatically and are estimates, not measurements. They are not medical advice. See Section 17 for the full AI transparency disclosure.
• Temporary Image Transmission: When you analyze a meal photo, the image is temporarily sent to OpenAI's Vision API for food recognition.
• Not Used for Training: Your photos are never used to train our models, and are sent to the AI provider solely to produce your result.
• Limited Retention: Images are sent to OpenAI solely for analysis and are not used to train AI models. OpenAI may retain API data temporarily (up to 30 days) for abuse monitoring before deletion. We do not store your meal photos on our own servers.
• Image Compression: Before transmission, images are compressed (WebP/JPEG, max 1024px, quality 0.8) to minimize data transfer.
• Secure Transmission: All API communications use encrypted HTTPS connections.

Purchases and Scan Tracking

• Subscription status is managed by RevenueCat and synced across devices.
• Scan pack balance (one-time purchases) is tracked locally and synced to Firestore.
• Daily free scan count is tracked and synced for cross-device consistency.

AVGusto can optionally connect to Apple Health (iOS) or Health Connect (Android) to give you a calorie budget that adapts to your physical activity. This connection is optional — the app works fully without it, and we ask for your permission before any health data is accessed.

Data We Read

• Workouts and exercise sessions (including type, duration, and distance), active energy burned, and step counts — used to adjust your daily calorie budget.
• Body weight — used to track your progress and to recalibrate your calorie targets over time.

Data We Write

• When you log a meal, AVGusto writes the nutrition summary (calories, protein, carbohydrates, fats) back to Apple Health / Health Connect so your "energy in vs. out" picture stays complete. On Android, deleting a meal also removes the corresponding Health Connect entry. On iOS, entries already written to Apple Health are not removed by the app — delete them in the Health app if you no longer want them.

How This Data Is Handled

• Apple Health and Health Connect remain the source of truth. We read activity data on-demand and do NOT copy your raw workout history to our servers.
• Only the aggregates needed to power Pro activity trends, along with your body-weight entries, are synced to Firebase Firestore for multi-device access.
• Health data is never used for advertising or shared with advertising partners.
• You can revoke health access at any time in your device's Health settings — the app continues to function without it.

Related Device Permissions

To connect Bluetooth fitness trackers and read motion data, AVGusto may request Bluetooth, Physical Activity (activity recognition), and Body Sensors permissions. On iOS and on older Android versions, the system may also request a location permission as a technical prerequisite for Bluetooth scanning — it is never used to collect or store your location.

AVGusto uses Google AdMob for rewarded video ads only. When your daily free scans are exhausted, you may choose to watch a rewarded video ad to earn one bonus scan.

• Opt-in only: Rewarded ads are never shown automatically — you choose to watch them.
• You control ad personalization: When you first sign in, you choose whether to see personalized or non-personalized ads, and you can change this at any time in Settings.
• Non-personalized ads rely only on contextual information and do not use your advertising identifier to track you across apps.
• Personalized ads allow Google AdMob to use your device's advertising identifier to tailor the ads you see, as described in Google's Privacy Policy. On iOS, personalized ads are shown only if you also grant permission through Apple's App Tracking Transparency prompt; if you decline, you will only see non-personalized ads.
• Pro subscribers and users with available scan packs are not shown ads.

• UTM Attribution: If you open the app via a deep link with UTM parameters, these are captured once (first-touch) and stored in Firestore for marketing analytics.
• Firebase Crashlytics: We collect anonymous crash reports to improve app stability. No personal data is included in crash reports.

• No location data
• No contacts or social connections
• No browsing history
• No cross-app tracking unless you opt into personalized ads (see Section 8); if you choose non-personalized ads, your advertising identifier is not used to track you
• No data sold to third parties

OpenAI Vision API

• Purpose: Food recognition and nutrition analysis
• Data Shared: Compressed meal photos (temporary processing only)
• Policy: OpenAI Privacy Policy

Firebase (Authentication, Firestore, Storage, Crashlytics)

• Purpose: User authentication, profile and data sync, avatar storage, crash reporting
• Data Shared: Email, display name, profile data, nutrition goals, scan counts, food-diary entries (food names and nutrition values, without photos), crash logs
• Policy: Firebase Privacy Policy

RevenueCat (Subscription and Purchase Management)

• Purpose: Managing subscriptions and one-time scan pack purchases
• Data Shared: Purchase history, subscription status, product identifiers
• Policy: RevenueCat Privacy Policy

Google AdMob (Rewarded Ads)

• Purpose: Serving opt-in rewarded video ads for bonus scans
• Data Shared: If you opt into personalized ads, your device's advertising identifier (as per Google's ad policies); if you choose non-personalized ads, no advertising identifier is shared for cross-app tracking
• Policy: Google Privacy Policy

USDA FoodData Central API

• Purpose: Nutritional data lookup for recognized food items
• Data Shared: Food search queries (no personal data)

Open Food Facts

• Purpose: Barcode product lookup and multilingual food search in the manual entry flow
• Data Shared: Barcodes scanned by you and text search queries you type (no personal data, no account identifiers)
• Policy: Open Food Facts Terms of Use

Apple Health & Health Connect

• Purpose: Reading activity, workouts, steps, and weight; writing nutrition summaries back (see Section 7)
• Data Shared: Processed on your device — health data is not transmitted to our servers beyond the aggregates described in Section 7

• Permission Control: You can grant or revoke camera and photo library permissions at any time through your device settings.
• Ad Personalization: You can switch between personalized and non-personalized ads at any time in Settings.
• Health Access: You can connect or disconnect Apple Health / Health Connect at any time from your device's Health settings; the app continues to work without it.
• Data Deletion: You can delete any meal entry, photo, or all data from within the app.
• Account Deletion: You may request deletion of your account at any time from within the app. This removes your Firestore profile, food-diary entries, body-weight and allowance data, and your avatar from Firebase Storage.
• Offline Usage: The app functions fully offline except for AI analysis, authentication, and sync features.

AVGusto is designed for users aged 16 and older. We do not knowingly collect personal information from anyone under 16. If you are a parent or guardian and believe your child has provided personal information, please contact us so we can promptly delete it.

We may update this Privacy Policy from time to time to reflect changes in our practices or legal requirements. We will notify you of any material changes through the app or by email.

If you have any questions about this Privacy Policy, please contact us:

• Email: [email protected]
• LinkedIn: linkedin.com/in/maksym-rakomin

• Meal photos stay on your device — never uploaded to our servers.
• Avatar photos you upload are stored securely on Firebase Storage.
• AI processing is temporary and stateless.
• Health data is read on-demand and stays in Apple Health / Health Connect — we never copy your workout history to our servers, and never use it for advertising.
• Rewarded ads are opt-in only, and you control whether they are personalized.
• You have full control over your data.
• No hidden data collection; cross-app tracking only ever happens with your explicit opt-in to personalized ads, never silently.

This section is provided in line with the transparency obligations of Regulation (EU) 2024/1689 (the EU AI Act).

Where AI Is Used

AVGusto uses artificial intelligence in one place: reading the photos you take of your meals. The AI recognizes the foods in the picture and estimates their portion size, calories, protein, carbohydrates and fats. Barcode scanning and manual entry do not use AI — those values come from food databases.

How It Works

• Food recognition runs on third-party AI vision models. We are the provider of the AVGusto AI system; we do not build or train the underlying models ourselves.
• Where a match is found, the AI's nutrition estimate is cross-checked against public food databases (USDA FoodData Central, Open Food Facts) before it is shown to you.
• Every AI-generated result is labelled as such in the app, on the screen where it appears.

Limits of the Output

• Values are estimates, not measurements. Accuracy depends on the photo, the dish and the portion; mixed dishes, sauces and ingredients that are not visible are the hardest to read.
• AVGusto is a general-purpose nutrition tracker. It is not a medical device, does not make medical decisions, and does not replace advice from a doctor or dietitian.
• You can edit or delete any AI-generated entry before and after it is saved.

Questions

Questions about how AVGusto uses AI can be sent to [email protected]. A plain-language explanation is also published at avgusto.app/ai.