Effective date: October 10, 2026
Link is operated by Hursh Desai ("we," "us," or "our"). This policy explains how we collect, use, store, and share information when you use Link. Contact us at hurshdesai8@gmail.com.
Account information. When you sign in, Clerk processes your account identifier and authentication information. Phone sign-in uses your phone number. If you choose username and password sign-in or sign-up, Clerk processes those credentials and any additional fields or verification information required by that flow. Sign in with Apple can supply an Apple account identifier, name and email address, including a private relay email. The information returned depends on Apple authorization and the enabled Clerk configuration. Link does not store your password in its library backend.
Your saved library. We store the URLs you save, link titles and previews, notes, tags, folder names, folder extraction instructions, and generated summaries and documents. For supported sources, processing may retrieve and store source text, images, video, and associated metadata. Your library is associated with your account.
Search and AI inputs. With AI enabled, Premium search sends your query after you pause typing to find related results. Folder suggestions use your folder name. Summary generation uses your folder instructions and retrieved source material. Generated summaries are also processed to make them searchable. Link does not maintain a separate search-history table; processing providers may retain request information under their policies.
Subscription information. If you subscribe to Link Premium, Apple processes your purchase. RevenueCat and our backend process subscription status, product identifiers, expiration information, and an account-associated identifier to verify access. Link does not collect your payment-card number.
Technical and support information. We keep synchronization and processing records, such as device identifiers used for synchronization, operation identifiers, timestamps, processing status, and error codes. The sync identifier is a random identifier created on the device. Extraction diagnostics store an error code, processing stage, attempt count, and time. They do not store source text, prompts, or media addresses. If you contact us, we receive the information you provide.
Link does not include a separate analytics or advertising SDK. Clerk, Convex, Google, Apify, Apple, and RevenueCat receive the network requests required for their services, which can include an IP address and ordinary request logs. The included RevenueCat SDK's privacy manifest reports purchase history for app functionality and states that the SDK is not used for tracking. Link passes your authenticated library identifier to RevenueCat so subscription access can be matched to your account. RevenueCat also provides subscription analytics and customer activity history.
We use information to authenticate your account, save and synchronize your library, organize and search your links, generate summaries and folder suggestions, verify subscription access, operate and troubleshoot Link, and respond to support requests.
Link does not show ads and does not sell personal information. The app does not use Apple's App Tracking Transparency prompt or a cross-app tracking SDK. Service providers may still log technical data while providing authentication, storage, AI processing, extraction, and subscription verification.
Link uses these providers for the functions described below:
| Provider | Function and information processed |
|---|---|
| Clerk | Account authentication and session management; authentication information and account identifiers |
| Convex | Cloud synchronization, backend processing, storage, and search; account-associated library content, retrieved media, processing records, and subscription access records |
| Google Gemini API | Folder suggestions, summaries, and search embeddings; relevant folder names/instructions, source text and media, generated summaries, and search query text |
| Apify and configured extraction services | Retrieve supported posts and their media; source URLs and retrieved content |
| Apple and RevenueCat | App Store purchases and subscription verification; purchase/subscription information and relevant account identifiers |
Loading a link preview or opening a source can contact the source website or its media hosts. Those services may receive ordinary network request information. If you use Link's share controls, the information you select is passed to the application or recipient you choose.
These providers process information under their own terms. Clerk stores the authentication account. Convex stores the synchronized library, retrieved media, generated documents, search embeddings, sync records, and subscription status. Apify and the enabled extraction services retrieve the source post from the URL you save. Apple and RevenueCat process the subscription. Google receives the folder name, folder instructions, source text and media, generated summaries, and search query text sent for the features described above.
Link uses a paid Google Gemini API project. Under Google’s paid-service terms, prompts and responses are not used to improve Google’s products. Google may retain limited safety and security logs and temporarily store or cache data where it or its agents operate. This does not mean immediate deletion or zero retention. Google’s terms are available at Google Gemini API terms.
Link stores library data and pending saves on your device and synchronizes account-associated data to its configured cloud backend. Requests to Clerk, Convex, Google, Apify, Apple, and RevenueCat use HTTPS. Link does not use end-to-end encryption: those providers can access the information they need to perform the functions in this policy. Their systems may be in the United States and in other countries where they operate. A saved link stays on your device if it has not finished synchronizing.
We keep an active library, including saved links, notes, tags, folders, retrieved source media, generated summaries, and search embeddings, for as long as the account exists. Sync records and extraction diagnostics are kept with that library. Clerk keeps the authentication account until it is deleted. Subscription status is kept so Link can verify access. Emails you send us are kept to respond and to keep a record of the request. Convex and the other providers may keep backups under their own practices. Link does not run a separate automatic backup purge.
Deleting a link removes it from the active library and from search results. The backend marks that record deleted. The marked record, retrieved media, generated documents, embeddings, and extraction history stay stored. They are not erased on a fixed schedule.
Account deletion is available through Settings > Account & Privacy > Delete account. This disables further processing and requests deletion of your authentication account, RevenueCat customer record, owned library records, retrieved source artifacts and stored media. The app clears its local account cache and signs out. Apple-linked accounts reauthenticate to revoke Apple authorization. Backend cleanup continues after the app closes and retries temporary service failures; completion can be delayed while a provider is unavailable. A minimal account-identifier marker is retained for seven days to prevent old sessions from recreating deleted data. Provider backups, security logs and records required by law may follow the providers’ retention policies; deletion does not promise immediate removal from those systems. You can also request access, correction or deletion at hurshdesai8@gmail.com. Signing out alone does not delete your cloud library.
Canceling a subscription is separate from deleting an account. Manage or cancel Link Premium through your Apple Account subscription settings.
You can edit your saved links, notes, tags, and folder instructions. You choose when to copy or share extracted information. Link requests Reminders access when you choose to add a reminder; you can revoke that permission in iOS Settings. Calendar actions use Apple's event editor so you can review the event before saving it. Opening a place in Maps passes the selected place information to the map application.
AI processing is off until you choose Allow AI processing on the AI & Privacy screen. You can review or disable it in Settings > Account & Privacy. Saving, organizing and synchronization remain available with AI off. Premium search still matches saved text; related results, automatic extraction and folder suggestions require consent. Disabling AI stops new requests and prevents unfinished results from being published. It cannot retract data already sent to a provider. Enabling AI later does not automatically resend older bookmarks; Retry or Refresh starts a new request.
Hursh Desai operates Link from the United States. Email hurshdesai8@gmail.com to request access, correction, deletion, or a copy of your information, or to object to or restrict a use described in this policy. We may ask for the phone number or Apple account on the Link account so we can locate it. We do not sell personal information. If the law where you live provides additional rights, including rights under the GDPR, UK GDPR, or a US state privacy law, use the same email address. We do not have a separate EU or UK representative.
Link is intended for adults aged 18 and older. Before using the library, you must confirm that you are at least 18. We store the time of that confirmation with your account; we do not collect a birth date for this check. Confirming your age does not enable AI processing. If you are under 18, do not use Link.
We may update this policy as Link changes. The revised policy will be posted at this page, hursh.xyz/privacy/link, with a new effective date. For a material change, we will update that date before the change applies.
Privacy contact: Hursh Desai, hurshdesai8@gmail.com.