Serena Reader Privacy Policy

Last updated: September 6, 2026

This policy describes the Serena Reader Android app and its Serena AI service. Documents, scans, and voice recordings are stored locally in the app. Certain features you choose to use need internet access and may send text or speech to a service provider, as explained below.

Documents and local app data

Serena Reader accesses documents you select, import, scan, or share with it, including PDF, EPUB, images, and text. Imported copies are saved in the app's private storage so you can reopen and read them. The app stores library information, reading progress, bookmarks, favorites, and preferences such as language and narrator settings locally. Importing a web address connects to that website to download the selected content.

Camera, scanning, and OCR

When you choose to scan a document, the Google Play services document scanner uses the camera to capture one or more pages. The app saves the resulting scanned document and can extract text using free local OCR. OCR text can be saved as a separate document linked to the original scan. If you separately choose AI OCR Pro and confirm its disclosure, the app may send the selected page image, local OCR text, language, and limited nearby-page context to Serena AI for correction. The image is rendered and recompressed without original photo metadata before upload. AI OCR Pro uses AI credits; declining it leaves local OCR available.

Microphone, personal recordings, and speech input

Microphone access is requested for recording your voice. Recordings and locally processed copies are kept in private app storage. You can play, rename, or delete them. Voice processing, including volume normalization and trimming silence, happens on the device. Recording a voice does not create an AI voice clone, and Serena Reader does not automatically upload these recordings to Serena AI. Exporting a recording through Android sharing sends it to the destination you choose.

Speech input uses the Android speech recognition provider. The app requests offline recognition where available, but the provider may process audio online. The app explains this before you proceed. Recognized text is returned to the app and may subsequently be used for translation or an AI question you choose to submit. The recognition provider's own privacy terms apply to its processing.

Reading aloud and translation

Reading aloud uses an installed Android text-to-speech engine. Offline voices process text on the device; a voice reported as requiring a network connection may send the text to its provider. The app asks for confirmation before using such a voice. Available voices and provider behavior depend on your device and engine.

Supported offline translations run on the device after the required language models have been downloaded. Model downloads require internet access. Where online translation is offered, the app uses the Serena AI service with your confirmation. It sends the text to be translated, not the original page image or your voice recording.

Serena AI and online processing

When you confirm an online AI request, the app sends the question or instruction, requested language, document title, selected document excerpts, and relevant recent conversation messages to the Serena AI HTTPS backend. These are processed through Cloudflare Workers AI to answer questions, summarize, explain, simplify, or translate text. A whole-document summary can send document text in successive sections together with an intermediate summary after you approve that operation. AI OCR Pro processes each confirmed page image and accompanying text in memory; the Worker code does not add the image or OCR content to its durable storage or application logs.

The app obtains a short-lived session token automatically; you do not need an AI account or to enter an API key. Version 3.0.7 sends a random app-generated installation ID when connecting to Serena AI. The Worker stores only its one-way hash and uses it to bind the session to a server-side AI credit account, issue a limited one-time FREE trial, and prevent the same backend identity from receiving that trial again. Requests are also subject to per-account and shared service limits, and online AI may be unavailable when credits or limits are exhausted or the service is disabled.

The Serena AI Worker code does not persist document content, questions, conversation history, or model answers in its database. It processes them to return the requested response. Its durable storage contains aggregate usage counters, time windows, and service enablement and request-reservation state for enforcing limits. These are not a stored document or conversation history. The deployed Worker has application observability logging disabled. This does not mean that Cloudflare has no infrastructure or security records: Cloudflare handles service data under its own terms and privacy practices. See Cloudflare Workers AI data usage and the Cloudflare Privacy Policy.

On-device libraries and service providers

Serena Reader uses Google ML Kit for supported on-device recognition, language identification, and translation. Google states that ML Kit processes input images and text on the device and does not send those inputs or results to Google. ML Kit SDKs do collect technical information such as device and app information, performance metrics, installation identifiers (and device identifiers for some downloaded components), and diagnostic usage data. Language identification and translation also report identified or configured languages for diagnostics. Translation uses Firebase installation and remote-configuration components for these technical purposes. Model and scanner component downloads also connect to Google services. See ML Kit terms and privacy information and ML Kit Android data disclosure. Other installed speech providers apply their own privacy practices when you use their online features.

Sharing and referral links

If you choose to share Serena Reader, the app can include a locally generated referral code in its Google Play link. It can read install-referrer information supplied by Google Play and store a pending referral locally. This is not uploaded to Serena AI by the current app. Sharing is performed through the Android share sheet to the destination you select.

Google Play purchases and AI credits

Beginning with version 3.0.7, Serena Reader includes infrastructure for an optional monthly or yearly Pro subscription and consumable Serena AI credit packs through Google Play Billing. Google Play supplies product availability and the localized price. The app sends the product ID, Google Play purchase token, package name, and a random app-generated installation ID to the Serene backend so a purchase can be verified before active subscription status or credits are granted. It does not request or use the Android Advertising ID.

The backend stores a one-way hash of the random installation ID, Pro entitlement status, whether the one-time trial was granted, total granted/spent/remaining AI credits, a bounded credit-change ledger, short-lived AI credit reservations, completed request identifiers used to prevent duplicate charging, and one-way purchase-token receipts used to prevent the same purchase from being granted twice. Credit entries contain the time, operation type, amount, and reason, but not document text. It does not store the raw purchase token after verification. Google Play purchase verification requires the backend to contact the Google Play Developer API. Billing remains inactive until the products and server verification are configured; the app must not show a fabricated price or grant an unverified purchase. A Pro subscription and AI credits are separate: Pro does not provide unlimited online AI.

Keeping and deleting data

Saved documents, OCR text, recordings, and settings remain locally until you remove them or clear the app's data. You can delete saved documents and OCR results from the library and delete private recordings from the recording controls. Original scans, extracted OCR documents, original recordings, and processed recording copies are separate saved items: delete each copy you no longer want. Deleting an imported copy does not delete the source file outside the app or copies you exported or shared. Some app preferences can remain until app storage is cleared.

AI conversation history is held in the current app session's memory, not saved as a permanent chat archive. Use the chat's clear control to clear that conversation; deleting a document is not the same as clearing its open chat. Android's Clear storage option or uninstalling the app removes its private local data. The app's Android backup rules exclude its files and preferences from cloud backup and device transfer.

The backend retains its aggregate operational and credit ledgers to enforce limits, trial eligibility, and idempotency across restarts. It has no saved document-content archive for a user to delete. Clearing app storage or reinstalling resets the locally held random installation ID. Because Serena Reader does not yet maintain signed-in user accounts, remaining credits cannot currently be recovered on a different installation, and a determined user may obtain a new backend identity after clearing data or reinstalling; stronger cross-install trial enforcement requires a privacy-reviewed Play or account identity mechanism. Provider-side retention is governed by the relevant provider's policies.

Data security and your choices

The app uses private Android storage for its saved documents and recordings, system document selection, and permission checks for microphone use. Online Serena AI communication uses HTTPS and short-lived session authorization. Server credentials remain on the backend and are not embedded in the app. No storage or transmission method can guarantee absolute security. You can decline online processing, use available offline features, revoke permissions in Android settings, and remove locally saved items.

Children's privacy

Serena Reader is intended for adults aged 18 and older and is not directed to children. The app does not request a date of birth or maintain user accounts; this audience designation is not an in-app age-verification system. If you are a parent or guardian with concerns about a child's use of the app or information submitted through it, please contact us using the privacy contact below.

Privacy contact and updates

For privacy questions or concerns, email support@serenereader.app.

Changes to this policy will be published on this page with an updated date.