Privacy
This document describes how the Narra app for iOS and Android handles data. The get-narra.app website policy is a separate section below.
In short
- There is no account: no email, no password, no social sign-in.
- We do not sell data or show ads. AppMetrica (Yandex) is used for usage statistics and crash diagnostics.
- The books you add are stored on your device. To analyse a book, Narra uploads it to its server. Retention and deletion are described below.
Who processes the data
The operator is Mikhail Naer. Contact: mikhail.naer@gmail.com.
App server: api.narra.disrupt.builders.
What the app sends to the server
Installation identifier
The app creates a random installation identifier and secret and keeps them in protected device storage. They provide server access and link requests from one installation without requiring a name, phone number or email. iOS Keychain records may survive uninstalling; uninstalling does not delete server data.
Book content
The books you add live in the app’s storage on your device. In addition:
- File upload. Book analysis sends the book file to the server. A content hash identifies an existing upload; another upload may be needed after the server copy is deleted.
- Retention. Private files and analysis results are subject to automatic server-side expiry cleanup. Accessing material extends retention. Removing a book from your device does not immediately delete its server copy.
- Fragments of text. When narrating, talking to a character, or generating portraits, covers and illustrations, the app sends the relevant fragment of text, or a request built from it, to the server. These requests are processed to produce the result and are not used to train models.
- Books from the built-in catalogue are public-domain works; their content is already on the server, so no upload is needed.
Technical telemetry
Narra and AppMetrica receive app usage and reading events, device and OS information, installation identifiers, crash reports and diagnostics. The recipient also sees the IP address; AppMetrica may use it to determine an approximate region. Book text, titles, character messages and access keys are not added to custom analytics events. The app has no separate analytics switch. Starting with Android build 21, advertising identifier collection is disabled. The main Narra server uses HTTPS.
Who else receives data
For the app to work, our contractors process some requests:
| Purpose | Processor |
|---|---|
| Speech synthesis (narrating books and characters) | speech synthesis providers connected to the Narra server |
| Text models (book analysis, conversation with a character) | language model providers connected to the Narra server |
| Image generation (portraits, covers, scenes) | image generation providers connected to the Narra server |
| App usage statistics and diagnostics | AppMetrica — YANDEX LLC |
| Hosting for the get-narra.app website | Vercel Inc. |
They receive only what the request requires. The access keys for these services are kept on the server and never reach the app.
If you connect your own AI service
In the settings you can supply your own key and endpoint for a third-party model, for translation or semantic search. Those requests then go directly to the provider you chose, and that provider’s policy applies to them instead of this one. Your key is kept in the device’s secure storage and is never sent to our server.
Device permissions
- Background audio — so that an audiobook keeps playing when the screen is off.
Storage and deletion
- On the device: books, covers, portraits, generated audio, reading progress, bookmarks and settings. All of it is removed together with the app.
- On the server, private books and analysis results expire after a retention period that is extended by access. Logs, installation identifiers, AppMetrica statistics and AI-provider data are stored separately; there is no single deletion period for all data.
- Any generated image or narration can be deleted or generated again from inside the app.
- To request deletion, email mikhail.naer@gmail.com. Include the platform, app version, approximate time of use and, for book data, the title and author. We will clarify the information needed to locate the data and the deletion process. Do not send passwords, access keys or the full book file.
Children
The app is intended for adults aged 18 and over. Store content ratings may differ by country. We deliberately do not collect data from children.
Content created by models
Portraits, illustrations, summaries and the replies of characters are created by language models and may be inaccurate or inappropriate. Requests pass through server-side moderation; material that may not be distributed is not allowed in. Any generation can be redone or deleted. Report an inappropriate result to mikhail.naer@gmail.com.
Changes
We will update the date at the top of this document when something material changes. The current version is always available at https://get-narra.app/en/privacy.
The get-narra.app website policy
The website uses no cookies, counters, analytics or advertising trackers. There are no forms, sign-ups or accounts. The hosting provider, Vercel, processes technical request data — your IP address, browser type, the page requested and the time of the request — to deliver pages and protect the site from abuse.
Last updated: