Privacy Policy
Effective date: May 25, 2026
Refluxora: AI GERD & LPR Companion ("the App") is developed and published by Valerii Loveiko. We take your privacy seriously and are committed to protecting your personal information. This privacy policy explains how the App collects, uses, stores, and protects your data.
Summary: Your reflux symptoms, RSI scores, meal photos, restaurant menu photos, medication logs, Acid Watcher diet adherence, taper protocols, and AI Coach chat history are stored locally on your device. Meal photos, menu photos, and Coach messages are sent to secure third-party AI services only after you give per-feature consent, and are not retained by those providers after processing. We do not collect your name or email, we do not use advertising or tracking SDKs, and we do not sell, share, or monetize your personal data.
Educational only: Refluxora is a Health & Fitness wellness companion, not a medical device. The App does not diagnose, treat, cure, or prevent GERD, LPR, Barrett's esophagus, or any condition. Always consult your doctor, gastroenterologist, otolaryngologist (ENT), or OB-GYN before starting, changing, or stopping any treatment. See the
Medical Disclaimer.
1. Information We Collect
1.1 Reflux & Symptom Data You Enter
Refluxora lets you record reflux-related health information so you can track it over time. This includes:
- Reflux symptoms — heartburn, regurgitation, throat clearing, hoarseness, globus (lump-in-throat), nighttime cough, post-meal burn — with severity and duration
- Reflux Symptom Index (RSI) scores — a 9-item Belafsky screening questionnaire you complete weekly for LPR (silent reflux) trend-watching
- Tums (calcium carbonate) and antacid dosage logs against the daily ceiling (relevant for Pregnant Patricia and Sam cohorts)
- PPI taper week and dose log (relevant for Anti-Pharma Anna cohort)
- Voquezna (vonoprazan) efficacy log (relevant for Burning Brian cohort with refractory GERD)
- Optional onboarding answers (age band, the cohort that best describes your situation: Burning Brian / Pregnant Patricia / Foodie Frank / Senior Sam / Anti-Pharma Anna / LPR Larry)
- Days-since-flare counter, vocal hygiene checklist, Acid Watcher diet phase adherence, supplement timing (DGL, zinc-carnosine, alginate)
All of this is stored exclusively on your device. It is used for trend charts, the doctor-shareable PDF (GI / ENT / OB / Primary-care formats), and to give the AI Coach context. It is never transmitted to any external server.
1.2 Meal Photos (Photo Food Scanner)
When you use the Meal Scan, the App accesses your camera or photo library at your explicit request to capture a meal. These photos are:
- Used to generate an educational reflux-trigger rating for the meal (a simple safe / watch / alert indication with a 0–10 trigger score)
- Stored locally on your device as part of your meal log
- Processed on-device by Apple Vision first; sent to OpenAI's vision model only as a fallback (after per-feature consent) when on-device confidence is low, then not retained by the provider
- Never used for face recognition or any biometric analysis — the scanner looks at food, not people. Refluxora does not compute, store, or transmit facial landmarks or biometric templates
The App does not scan your photo library and does not collect EXIF metadata (location, date, camera model) from your photos.
1.3 Menu Photos (Menu Pre-Scan)
Refluxora's Menu Pre-Scan lets you photograph a restaurant menu before ordering so the App can rank dishes for your cohort. These photos are:
- Used to extract dish names via Apple Vision OCR (runs on-device), then ranked per cohort scoring (refractory triggers / FDA Category-B safety / flavor-preserving swaps / drug interactions / Acid Watcher phase / pH score)
- Stored locally on your device only as a "Menu Pick" entry; the photo itself is discarded after OCR unless you keep it manually
- Sent to OpenAI GPT-4o-mini vision fallback ONLY when on-device OCR confidence is low (after per-feature consent), then not retained by the provider
- Never used for restaurant tracking, location, or any identification of you or the venue
1.4 AI Reflux Coach Chat Data
When you use the AI Coach, your messages and the AI's responses are handled as follows:
- Messages are sent — only after explicit per-feature consent — via encrypted HTTPS to our AI text provider (see §3.2), proxied through our Cloudflare Worker
- Chat history is stored locally on your device
- Every Coach response is grounded in an on-device clinical knowledge base; citations link to the source guideline (ACG 2022, AGA 2024, AAO-HNS LPR position paper, Mayo Clinic, Dr. Koufman and Dr. Aviv Acid Watcher protocols, Voquezna FDA label, ACOG pregnancy guidance, IFFGD lifestyle, AGS Beers Criteria, and selected PubMed extracts)
- The on-device knowledge base never leaves your device
- No photos, lab-report images, identifiers, or Apple Health records are sent to the chat provider — only your message text and a brief on-device context summary (your cohort, age band, recent symptom severity, and the last meal or menu pick if relevant) so the answer fits your situation
1.5 Medication and Protocol Logs
If you log medications (for example PPI dose, Voquezna trial, H2 blockers, Tums adherence, alginate timing), supplements (DGL, zinc-carnosine, mastic gum, melatonin), or Acid Watcher diet phase progress, this information is:
- Stored exclusively on your device
- Used for timeline visualizations, taper tracking, drug-interaction flags (Sam cohort — grapefruit + amlodipine etc.), and the doctor-shareable PDF
- Never transmitted to any external server
Refluxora does not prescribe, supply, or modify any medication or dose. The medication log is solely a personal journal of what you are already doing under your own clinician's care.
1.6 Apple Health Data
With your permission, Refluxora can read selected data from Apple Health (for example sleep position and duration, dietary water intake, body mass). This access is read-only — the App does not write to Apple Health. Data read from Apple Health is used on-device for sleep-position + meal-timing correlation and never transmitted off the device. You can grant or revoke this access at any time in iOS Settings → Privacy & Security → Health.
1.7 Subscription Information
Refluxora offers optional subscriptions managed entirely through Apple's App Store and StoreKit. All payment processing and billing are handled by Apple — we do not collect, process, or store payment information. We use RevenueCat to track subscription status (see §3.4).
1.8 Anonymous Analytics
To understand aggregate App usage and reliability, Refluxora records a small amount of anonymous event data (for example "onboarding completed", "paywall viewed", "Menu Pre-Scan completed", a subscription event from RevenueCat) in our Supabase backend (see §3.5). These events are tied only to an anonymous, system-generated identifier. They contain no name, no email, no symptom values, no photos, and no chat content.
1.9 Information We Do NOT Collect
Refluxora does not collect:
- Your name, email address, phone number, or any contact information
- Your physical location or GPS coordinates
- Restaurant names, addresses, or any venue identification
- Advertising identifiers (IDFA) or any identifier used for cross-app tracking
- Browsing history or activity outside the App
- Contacts, calendar, or other personal data from your device
- Biometric identifiers, face-recognition templates, or face embeddings
- Social media accounts or login credentials
2. How We Use Your Information
The information processed by the App is used exclusively for:
- Today tab cohort-routed dashboard: local visualization of your reflux markers, RSI trend, Voquezna efficacy, or taper-week progress per cohort
- Meal Scan: educational reflux-trigger rating for a meal you photograph
- Menu Pre-Scan: per-cohort ranking of dishes from a restaurant menu photo before ordering
- AI Reflux Coach: generating sourced, citation-backed educational answers about GERD, LPR, pregnancy reflux, polypharmacy interactions, taper protocols, and Acid Watcher diet
- Medication & Protocol tracking: local logs for PPI / Voquezna / Tums / taper / Acid Watcher adherence
- Doctor-shareable PDF: on-device export of your trend data formatted for your GI, ENT, OB, or Primary-care appointment
- Personalization: adapting in-app copy, Coach prompts, paywall hero, and ranking algorithms to the cohort you selected
- Reliability: anonymous, aggregate analytics to fix bugs and improve the App
We do not use your data for advertising, profiling, or any purpose unrelated to the core functionality of the App.
3. Third-Party Services
3.1 OpenAI (Meal & Menu Vision OCR Fallback — Optional)
When on-device Apple Vision OCR has low confidence reading a meal or menu, Refluxora may use OpenAI's GPT-4o-mini vision model as a fallback. This call only happens when you explicitly request the scan, when local OCR is insufficient, and after you have accepted the per-feature consent prompt.
- The image is transmitted via encrypted HTTPS to OpenAI's API, proxied via a Cloudflare Worker
- OpenAI returns recognised dish names or a reflux-trigger rating
- API inputs and outputs are not used to train OpenAI's models and are not retained after processing (zero-retention API usage)
- No personal identifiers are sent alongside the image — only the image and the analysis instruction
- OpenAI's `text-embedding-3-small` model is also used to embed your Coach question text for retrieval against the on-device knowledge base; the question text is sent zero-retention; results stay on-device
- For details, see OpenAI's API Data Usage Policy
3.2 DeepSeek (AI Reflux Coach Text)
Refluxora uses DeepSeek V3's API to generate AI Coach responses, grounded by our on-device reflux knowledge base. When you send a message:
- The message text is transmitted via encrypted HTTPS to DeepSeek's API, proxied via a Cloudflare Worker
- DeepSeek returns a generated response that the App grounds against the on-device citation corpus before presenting it
- No photos, lab images, identifiers, or Apple Health data are ever sent to DeepSeek
- A brief context summary (cohort, age band, recent symptom severity, last meal or menu pick if relevant) is included so the response is tailored to your case
- API inputs are not retained after processing
3.3 Cloudflare (API Gateway)
All AI requests are routed through a Cloudflare Worker proxy (refluxora-api.veribag.workers.dev) that forwards traffic to OpenAI and DeepSeek without logging or storing request bodies. The Worker exists solely to keep API keys server-side, to apply rate limiting, and to provide IP anonymization between your device and the AI providers.
3.4 RevenueCat (Subscription Management)
Refluxora uses RevenueCat to manage subscription status. RevenueCat receives an anonymous, system-generated user identifier, your subscription status (active, trial, expired), and the product identifier purchased. RevenueCat does not receive your name, email, symptoms, photos, logs, or any personal information. For details, see RevenueCat's Privacy Policy.
3.5 Supabase (Anonymous Analytics Backend)
Refluxora uses Supabase to store anonymous usage and subscription events (see §1.8). Each event is tied only to an anonymous identifier and contains no personal health data, photos, or chat content. Supabase acts as our data processor for this limited, anonymous telemetry. For details, see Supabase's Privacy Policy.
3.6 Apple Services
- StoreKit 2: subscription management and payment processing. All financial data is handled by Apple.
- HealthKit: read-only access to the Apple Health categories you approve (sleep, dietary water, body mass). Apple Health data stays on-device.
- Vision: on-device OCR for menu and meal text/dish recognition. No data leaves the device for Vision processing.
- iCloud Backup: if enabled on your device, local App data may be included in your device backup. This is controlled by your device settings.
3.7 No Advertising or Third-Party Tracking SDKs
Refluxora does not integrate advertising networks, cross-app tracking SDKs, or third-party behavioral-analytics platforms (Google Analytics, Firebase Analytics, Mixpanel, Amplitude, the Meta SDK, and similar). The App contains no advertisements. The only off-device data flows are the AI processing, RevenueCat, and the anonymous Supabase telemetry described above.
4. Data Storage and Security
4.1 Local Storage
All reflux symptoms, RSI scores, meal photos, menu picks, medication logs, taper history, chat history, symptom notes, and personalization settings are stored locally on your device using Apple's standard storage frameworks, with file protection enabled. This data resides only on your device, is protected by your device's passcode / Face ID / Touch ID and encryption, is not accessible to the developer, and can be deleted at any time from within the App or by deleting the App.
4.2 On-Device Knowledge Base
The reflux clinical citation corpus ships embedded in the App (~65MB sqlite-vec database, ~2,400 chunks) and is queried entirely on-device. It contains extracts from the American College of Gastroenterology (ACG 2022 GERD guideline), American Gastroenterological Association (AGA 2024 Best Practice), American Academy of Otolaryngology–Head and Neck Surgery (AAO-HNS LPR position paper), Mayo Clinic and Cleveland Clinic patient material, Dr. Jamie Koufman and Dr. Jonathan Aviv Acid Watcher protocols, the Voquezna (vonoprazan) FDA label, ACOG pregnancy reflux guidance, IFFGD lifestyle material, AGS Beers Criteria for senior polypharmacy, and selected PubMed literature on GERD, LPR, refractory reflux, and PPI taper protocols. It never leaves your device and is not personalized.
4.3 Network Security
All network communications are encrypted using industry-standard TLS 1.2+/HTTPS. No data is transmitted in plain text. The App enforces Apple's App Transport Security.
4.4 API Key Security
Third-party API credentials are kept server-side on the Cloudflare Worker and are not stored in plain text within the application.
5. Data Retention
- Reflux symptoms, RSI scores, photos, logs, chat history: retained on your device until you delete individual items or uninstall the App
- OpenAI / DeepSeek requests: not retained after processing — zero-retention API usage
- Cloudflare Worker: no request bodies logged or retained
- RevenueCat: anonymous subscription records retained per their data-retention policy
- Supabase: anonymous event records retained for aggregate analytics; they contain no personal data
- Subscription data (Apple): managed entirely by Apple
6. Your Rights and Choices
6.1 Camera, Photo Library and Health Access
You can revoke the App's camera, photo-library, or Apple Health access at any time through iOS Settings → Refluxora (and iOS Settings → Privacy & Security → Health). Without these permissions you can still use Refluxora as a manual tracker.
6.2 AI Consent
The first AI Coach message, the first Meal Scan, and the first Menu Pre-Scan each prompt you for explicit, separate consent. You can decline any of them and continue using Refluxora as a manual tracker. The post-onboarding Disclaimer screen also requires a separate "I consent to AI-assisted insights" checkbox before any AI feature is enabled. You can re-show these prompts from Settings.
6.3 Deleting Your Data
Because your personal data is stored locally on your device, you can delete individual entries within the App, use Settings → Delete all my data to wipe every log at once, or delete all data by uninstalling the App. No request to the developer is necessary — we do not hold your health data on external servers.
6.4 Subscription Management
Manage, cancel, or modify your subscription through iOS Settings → Apple ID → Subscriptions. Cancellation takes effect at the end of the current billing period. Refunds are handled by Apple.
7. Children's Privacy
Refluxora is not directed at children under 12 and is rated 12+ on the App Store. We do not knowingly collect personal information from children. If a parent or guardian is tracking on behalf of a younger family member, the consenting adult enters and reviews all data — no separate account or personal data is collected from the child.
8. International Users and Compliance
Refluxora is available in five launch markets (United States, United Kingdom, Canada, Australia, Germany) and complies with applicable data-protection regulations:
- GDPR (EU/EEA) & DSGVO (Germany): minimal data collection, no profiling, no advertising. Legal basis: your consent (camera, photo, Health and per-feature AI consent) and legitimate interest (core service and anonymous reliability telemetry). Right to erasure: delete your data in-App or uninstall the App.
- UK GDPR: the same rights and protections apply.
- CCPA/CPRA (California): we do not sell or share personal information and do not use it for behavioral advertising.
- PIPEDA (Canada): compliant — minimal, consent-based, secure.
- Australian Privacy Act: compliant — privacy-by-design, local storage, no tracking.
9. Medical Disclaimer
Refluxora is an educational wellness companion, not a medical device. It does not diagnose, treat, cure, or prevent GERD, LPR, Barrett's esophagus, or any condition. Symptom tracking, RSI scores (Reflux Symptom Index — a screening tool, not diagnostic), meal trigger ratings, menu pre-scan rankings, and AI Coach answers are educational information and self-reported tracking, not clinical assessments. Always consult your doctor, gastroenterologist, otolaryngologist, or OB-GYN. See the full Medical Disclaimer.
10. Changes to This Policy
We may update this policy to reflect changes in functionality or regulations. Material changes will be indicated by updating the effective date. Continued use after changes constitutes acceptance of the updated policy.
11. Contact Us
For questions about this privacy policy or your data:
Medical Disclaimer ·
Terms of Use ·
Support