Privacy Policy for AReader

Last updated: July 9, 2026

AReader (“the app”) is developed by Vladimir Vsevolodovich (“we”, “us”). This policy explains what data the app collects, why, and how it’s handled.

If you have questions, contact vladimir@supplied.eu.

1. Data that stays on your device

The app is designed to keep your library and reading data local wherever possible:

• Books you import (EPUB/PDF files) and any covers extracted from them
• Reading progress, bookmarks, and high
• AI chat history (your questions and the answers you received)
• On-device AI search index — when you import a book, the app computes text embeddings locally (EmbeddingGemma model) and stores them in a local vector index (sqlite-vec). This never leaves your device.

None of the above is uploaded to us or any third party. It’s stored in the app’s local database and deleted if you uninstall the app or clear its data.

The app requests access to your device’s media/files storage only so you can pick EPUB/PDF files (and, where applicable, cover images) to import — we do not access, scan, or upload your photos or other media.

2. Data sent to provide AI Chat

When you ask the in-book AI assistant a question, the app sends the following to a backend we operate (a Cloudflare Worker), so it can generate an answer:

• The book’s title and author
• Short excerpts from the book relevant to youry from content up to your current position in thebook, never beyond it (spoiler guard)
• Your question and, for context, your last fewsation
• An anonymous subscriber identifier (issued by our subscription provider, RevenueCat) used only to enforce free-tier usage limits and verify paid access — this identifier is not linny other personal information, because the app does nothave user accounts or logins

That request is processed by our Cloudflare Worker, which — depending on whether you’re on the free tier or a paid “Enhanced” plan — forwards it to onnerate the answer:

• Cloudflare Workers AI (Meta rd-tier requests
• Anthropic’s Claude (via Cloudflare AI Gateway), for Enhanced-tier requests

We do not store the content of your questions orthem on our servers beyond what Cloudflare’s AIGateway retains for request logging and cost tracking. We do not use this content to train AI models.

3. Subscriptions

Purchases and subscription status are handled by RevenueCat. The app shares an anonymous app-user identifier and purchase/entitlement events with RevenueCat and, through it, with Google Play (which processes your actual payment — we never see your payment details). See RevenueCat’s privacy policy for details on what they collect.

Our backend also queries RevenueCat with that same anonymous identifier to verify your entitlement (free vs. Standard vs. Enhanced) before answering an AI request.

4. Model downloads

The first time you use on-device AI features, the app downloads an embedding model file over the internet. This is a one-way file download — no personal or book data is sent as part of it.

What we don’t do

• We don’t require or offer account creation/sign-in at this time — the app has no way to identify you personally.
• We don’t run analytics or crash-reporting SDKs.
• We don’t show ads or share data with advertisers.
• We don’t sell your data.

6. Data retention and deletion

• Local data (library, highlights, progress, chat history, search index) is deleted when you uninstall the app or clear its storage from Android settings.
• Server-side, we retain only: a per-device free-quota counter (keyed to the anonymous subscriber identifier) and standard request logs kept by Cloudflare’s AI Gateway for cost t or resetting your subscriber identifier effectively starts these over.
• We do not retain book content or chat transcripts on our servers.

7. Third-party processors

Provider	Purpose
Cloudflare (Workers, AI Gateway, Workers AI, KV storage)	Runs the AI chat backend and enforces usage limits
Anthropic (Claude)	Generates answers for Enhanced-tier AI chat
RevenueCat	Manages subscriptions and entitlements
Google Play	App distribution and payment processing

8. Children’s privacy

AReader is not directed at children under 13, and we do not knowingly collect personal information from children.

9. Your rights

Because the app doesn’t collect personal iity (no accounts, no name/email), most data subject rights are exercised directly on your device: uninstalling the app or clearing its data removes your local library, highlights, and chat history. If you have questions about the limited server-side dder GDPR), contact us at vladimir@supplied.eu andwe’ll assist.

10. Changes to this policy

We may update this policy as the app evolves (e.ud sync is introduced in a future version). We’llupdate the “Last updated” date above when we do.