Hearleaf Privacy Policy

Effective date: August 2, 2026. This policy explains how Hearleaf handles information in the Hearleaf mobile app and the cloud services that support page preparation, voice authorization, authentication, quota, and abuse prevention.

Short Version

Hearleaf turns page images into private, cached audio. Voice recordings, photographed pages, extracted text, and generated audio are personal data. The app stores the working library primarily on your device and sends only the data needed for requested cloud processing.

We do not sell your personal data. We do not use your voice recordings, page images, extracted text, or generated audio for advertising.

Who We Are

Hearleaf is a mobile reading application that helps people photograph printed pages and hear those pages read aloud as AI-generated speech. The developer named in the app store listing is Hearleaf.

Privacy questions and deletion requests can be sent to palande.prathmesh@gmail.com. You can also use the developer contact channel shown in the Hearleaf Google Play listing.

Information We Access, Collect, Or Process

Depending on how you use the app, Hearleaf may access, collect, or process the following information:

Voice recordings, including consent recordings and voice samples you record in the app.
Book page photos, selected page images, and optional local cover images.
Extracted OCR text and compact OCR metadata used to prepare audio. The app shows canonical detected text in a non-editable review step before generation; technical OCR diagnostics remain internal.
Generated AI audio files and playback progress for pages and voices.
Local library information such as book titles, authors, page order, bookmarks, and selected voice choices.
Firebase authentication identifiers, including an anonymous Firebase user ID and, if you choose Google sign-in, account identifiers such as your Google-linked Firebase ID and basic profile information provided through Firebase Authentication.
App authenticity, device, request, quota, queue, error, diagnostic, and service-log metadata used to operate and protect the service.
Pseudonymous measurement keys derived from successful page-job IDs and Firebase authentication identifiers, together with aggregate successful-page and processing-time measurements. The product-metrics dataset does not store raw Firebase IDs, page or voice content, playback, or listening activity.
A coarse country/region code from the device locale when a page-preparation job is submitted. Hearleaf uses this for delayed aggregate status only and does not store raw locale strings, timezone values, or IP-derived location for this feature.

App Permissions

Hearleaf requests microphone access so you can record short clips for a local voice profile. It requests camera access so you can photograph pages and optional cover images. Android builds may also request vibration permission for local touch feedback.

Hearleaf does not use microphone or camera permissions for background advertising, background tracking, or unrelated profiling.

How We Use Information

We use information to provide and protect Hearleaf, including to:

Create and manage local voice profiles and reading libraries.
Check whether a short voice recording is usable for the app.
Extract text from a page image and synthesize page audio in the selected voice.
Cache generated audio locally so you can replay it without regenerating it.
Authenticate users, apply App Check, enforce page-generation quota, and prevent abuse.
Debug errors, monitor reliability, and operate cloud infrastructure.
Measure, for owner-only product analysis, successful page processing, exact submit-to-success timing, and how many resolved Firebase identities have completed at least one page. These measurements are not used for advertising.
Show delayed aggregate page-processing timing and country/region mix on the public status page without exposing individual user, page, voice, request, or artifact details.
Respond to privacy, deletion, security, and support requests.

Where Information Is Stored And Processed

Your voice profiles, recordings, book library, page images, generated audio, playback progress, and related app state are primarily stored locally on your device.

Some features require cloud processing. Voice authorization sends the one-time challenge recording to Hearleaf's private GPU processor for blind phrase verification. Page text detection runs on your device and standard page-audio preparation sends the reviewed text plus the selected voice reference through a secured gateway. If you explicitly choose Improve text detection, Hearleaf also sends that page image through the secured gateway for private hosted OCR and replacement audio preparation. These requests are transmitted over HTTPS.

Hearleaf uses Google Cloud and Firebase as service providers for authentication, App Check, quota, page-job queueing, private temporary storage, service logs, and hosted OCR and speech-synthesis processing. When product measurement is enabled, Firestore also stores pseudonymous event receipts, identity membership, and aggregate counters that are available only through an owner-local console.

Public page-processing status uses delayed, rounded, sanitized job metadata. Country or region appears only as aggregate buckets after enough completed jobs share the same coarse device-region code.

We do not sell personal or sensitive user data. We share information only in limited ways needed to operate the app:

With service providers such as Google Cloud and Firebase, which process data on our behalf for app functionality, security, quota, and infrastructure operation.
On the public status page, as delayed aggregate timing, status, and country/region buckets that meet the minimum count threshold. Individual page jobs do not show country or region.
When you choose to sign in with Google, with Google/Firebase as needed to authenticate and link the account.
If required to comply with law, enforce rights, protect users, investigate abuse, or respond to valid legal process.
As part of a merger, acquisition, financing, or transfer of assets, with appropriate notice where legally required.

Retention And Deletion

Local app data remains on your device until you delete it in the app, clear app storage, or uninstall Hearleaf. This includes local profiles, recordings, book and page metadata, page images, and cached generated audio.

Voice-challenge recordings are processed in request-scoped temporary storage by the private GPU service and are not intentionally kept after the request finishes.

Page-preparation request artifacts are private temporary objects. Request artifacts are normally deleted after a job succeeds or reaches terminal failure. Result artifacts are normally deleted after the app downloads the result and acknowledges it. If a job does not reach a terminal state or result acknowledgement is missed, its private artifacts become eligible for asynchronous lifecycle deletion after 7 days; the cloud provider may complete that deletion later.

Provider-controlled deletion and shared-storage garbage collection can complete after Hearleaf removes an object from its active namespace. Ordinary Hearleaf users cannot access provider-restricted recoverable copies.

Firebase, Firestore, Pub/Sub, Cloud Run, and other operational systems may retain authentication identifiers, quota records, queue metadata, request metadata, service logs, and error logs according to their configured retention behavior. Current quota and usage records are designed to expire after about 90 days.

Product-metric event receipts and aggregate counts may be retained so successful-page totals and exact processing-time measurements can be reconstructed. Versioned one-way Firebase identity membership is retained until the related account-deletion workflow or an explicit metric reset. Deletion removes that linkable membership while preserving only its already-counted aggregate contribution; a later new Firebase identity may count as a new identity instance.

To request deletion of server-side account, quota, or operational metadata associated with your Firebase user, contact palande.prathmesh@gmail.com. Some records may be retained where needed for security, fraud prevention, legal compliance, dispute resolution, or infrastructure integrity.

Voice Consent And AI-Generated Audio

Voice recordings are sensitive. Hearleaf is intended for your own voice, or for a voice you have explicit permission to use. Do not record or use another person's voice without their consent.

Generated speech is synthetic AI audio. If you share generated audio outside the app, you should disclose that it is AI-generated and only share audio when you have the right to share the underlying voice and reading material.

Children

Hearleaf is not currently designed as a child-directed app. If child-directed or family account features are added later, this policy and the app's consent, safety, and deletion controls will be updated before those features are offered.

Security

Hearleaf uses HTTPS, Firebase Authentication, Firebase App Check, a secured gateway, quota checks, private Cloud Run services, and private temporary artifact storage to reduce unnecessary exposure of sensitive voice and page data. No app or internet service can guarantee perfect security, but Hearleaf is designed to minimize server-side retention of voice recordings, page images, extracted text, and generated audio.

Changes To This Policy

We may update this policy as Hearleaf changes. Material updates will be reflected by a new effective date and revised policy text at this URL.