Privacy policy

Effective November 14, 2026. Shelfmark is a personal home-library app provided by Pars Kunt, Muğla, Türkiye. Contact the controller using the support email below.

On your device

Your library is stored in SQLite on your device: works, editions, copies, locations, tags, wishes, reading dates and status, ratings, notes, purchase details and lending records. Search runs locally. Camera captures, original cover and copyright-page photos, and processed cover files are kept in the app’s local files. Preferences and sign-in session data use local app storage. Basic library browsing, manual entry and checking owned books work offline without an account. Device or operating-system backups are subject to your platform settings.

Optional accounts and cloud backup

Email sign-in sends your email to Supabase Auth. Apple and Google sign-in send identity tokens to Supabase; email and profile names supplied by those providers may be stored with the account. Apple’s name is saved as a display name when available. When signed in, Shelfmark backs up and syncs private library rows, including notes, collections, locations, wishes, reading and purchase details, lending names and dates, and extracted edition metadata. Only include other people’s details when you are entitled to do so. Processed custom cover JPEGs are uploaded to the private Supabase covers bucket. Original camera captures and the copy’s photo manifest are excluded from cloud backup. Access policies associate private rows and cover objects with your user ID. The current app starts sync for signed-in accounts without a Pro check.

Regions and technical requests

Cloud data is hosted in the region configured for our Supabase project; the repository does not specify that hosting region. We do not claim a particular country of storage. Supabase, Anthropic, Apple, Google and RevenueCat may process requests in other countries. The app uses the device’s language, region and currency settings locally for formatting and defaults; it does not upload a separate device-region profile field to Supabase. Store services can process storefront country and currency. Online requests expose an IP address and ordinary network information to the receiving service and its infrastructure.

AI photo reading

AI reading is optional and needs a signed-in account and internet in distributed builds. Cover and copyright/title-page JPEGs, including a smaller cover image used to find its corners, go to our authenticated Supabase extract-book Edge Function, then to Anthropic’s API for structured text extraction and cover positioning. Photograph only the book and avoid personal material. The function does not write image bytes to database or object storage; it forwards them during the request. The extracted text and cover geometry are cached in book_extraction_requests with your user ID, request ID and month. Monthly successful-use counts enforce the free quota. Successful cached results have no automatic deletion schedule in the current code and remain until account deletion or server-side removal. The two-minute expiry applies to pending reservations, not to completed results. Anthropic processes the submitted content under its applicable API terms and retention policies; we do not promise zero retention by that provider. A separate search-edition function can send title, authors and language to Anthropic to find edition details. Check AI results before saving. A direct Anthropic API path exists only in development when a developer supplies a key.

Book lookups and cover downloads

ISBN lookup checks your local library and our shared Supabase editions catalogue, then Google Books and Open Library. Photo enrichment can query those services by title, author, language or ISBN. They receive the search terms and ordinary request metadata, not your full private library. Remote cover URLs are downloaded from the returned image hosts. During cloud sync, confirmed ISBN edition metadata may be contributed to the shared catalogue: bibliographic facts such as title, authors, publisher, translator, language, year, printing, pages, genre and a public HTTPS cover URL. Private notes, shelf locations, copy details, account identifiers and original photos are not part of that public contribution.

Purchases

Apple’s App Store processes payment. We do not receive your payment-card details. RevenueCat manages Pro entitlements and restoration using transaction and purchase history, app user identifiers, and SDK device and request information. Signed-in Supabase user IDs are linked to RevenueCat; purchases without sign-in can use its anonymous app user ID. The installed SDK’s automatic device-identifier collection is enabled by default, while optional diagnostics are disabled by default. Our server can check RevenueCat and receives purchase webhooks to maintain Pro status. Storefront country and currency can be processed by purchase services. These uses support purchases, restoration and entitlement checks, not advertising.

No advertising or cross-app tracking

Shelfmark has no ads, advertising integrations, third-party analytics SDK, or app behaviour analytics events. We do not sell personal data or use it to track you across other companies’ apps and websites. Purchase management, account authentication, AI quota counts and normal service operation still process the data described above. This website uses Google Fonts, which receives ordinary font-request metadata, but no analytics, cookies, contact-form backend or advertising scripts. Your browser may cache fonts and static files.

Retention, export and deletion

Local data remains until you delete it or remove the app, subject to your device backups. Cloud library rows, uploaded covers, account data, quota records and cached extraction results are retained while your account exists; sync deletion markers can remain to propagate changes. In You, Delete account calls our delete-account function: it removes private uploaded covers, deletes the Supabase auth user and cascades private cloud records. Local library books are deliberately retained. Delete them separately if desired. Public shared bibliographic contributions do not identify you and can remain. App Store and RevenueCat transaction records are handled under their applicable retention obligations; deleting a Shelfmark account does not cancel a store subscription. Pro users can export library details as a UTF-8 CSV from You using the system share sheet. The CSV contains no image files. You can also contact us about access, correction, deletion or a data copy; rights requests are not conditional on buying Pro.

Permissions, support and your choices

Camera access is used for book scanning and photos. Photo-library selection, where used, is limited to the images you choose through the operating system. Denying camera permission leaves manual entry available. You choose whether to sign in, request AI extraction, buy Pro, export a file or email support. Support email contains the information you send, plus editable version and device details in the draft; we use it to answer your request. Contact Pars Kunt at the email below for privacy questions or to exercise applicable data rights. We may update this policy and will change its effective date when we do.