Privacy Policy

Controller: Rakhat Zhumaev, an individual developer who publishes Sparkhop
Contact for privacy requests: privacy@sparkhop.app
General support: support@sparkhop.app
Website: https://sparkhop.app

This Privacy Policy explains how Rakhat Zhumaev (“we”, “us”, “our”) collects, uses, stores and shares information when you use Sparkhop (the “App”) and related services. Sparkhop is published by an individual developer; there is no company behind it yet. If the service moves to a company, we will tell you in the App before the change and update this policy.

1. Information we collect

We collect the following categories of data.

• Account data: name, username, e-mail address, profile photo, and the identifiers issued by the sign-in method you choose (Apple or Google; GitHub is used only in test builds and by our administrators).
• Terms acceptance record: when you create an account we record the version of the Terms of Use you accepted and the time. We keep this as proof of the contract between you and us, for as long as your account exists.
• AI consent record: if you allow us to send pages to Anthropic’s Claude (section 3), the version of the text you agreed to and the time. We keep it as proof of your choice while your permission is on, and delete it when you turn it off or delete your account.
• Authentication data: session records and device/session tokens. We do not create sessions for visitors who are not signed in (see “Visitors without an account” below).
• Content you create: lists, wishes (items) with their descriptions, tags, photos you upload for your own wishes (a Plus feature, where available), likes, Quick Adds, event dates (“For an event” lists), shares and invitations, and your settings.
• Purchase marks: when you mark a wish on someone else’s list as reserved or bought, we store the mark, its time and your choice of whether the list owner may see your name. By default list owners do not learn that a wish was bought or by whom.
• Family and shared lists: membership of a family group (up to 6 people) and the people you share a list with, with their roles.
• Price tracking: the product pages of wishes you track and the prices found on them over time.
• Links you import: when you import a wishlist from a link, the page address and, preferably, the HTML of that page as rendered on your device (see section 3).
• Notification data: notifications shown in the App, your notification preferences, your time zone and, when push notifications are available and you turn them on, a push token for your device.
• Subscription data: your Plus or Plus Family subscription status, plan, store, period dates and an app-specific billing identifier (a random UUID generated by us). We receive this from the app store and our subscription provider; we do not receive your card number or store account details.
• Reports and moderation data: reports you file about content or people, and the automated moderation status of content you submit, with the reason category.
• Problem reports and ideas: when you send a problem report or an idea from the App (for example by shaking your phone, from Settings, or from an error screen), we receive what you write (up to 4,000 characters), whether it is a problem or an idea, and where in the App you opened the window. A problem report also carries a snapshot of the state of the App: app version, build and update; operating system version, device model, memory and low-power mode; screen and window size, text size, light or dark mode, theme and orientation (and the fold state of a foldable device); App language, device language and time zone; the screen you were on and the last few screens you visited (names of screens, never their content); whether you were online and the type of connection; the last few technical error events and failed requests (the kind of request, the name of the screen route, the error code and a request number, never the content); the state of the App’s sync queue; which App features are switched on; whether notification, photo and camera access is allowed; a rough band of free storage; and the random installation identifier of our diagnostics provider (section 4). If you are signed in, the report is linked to your account and may include a screenshot of the screen (the App attaches it by default, shows you a preview, lets you cover private details and lets you remove it with one tap) and one photo you choose from your library. A screenshot or photo may show whatever was on the screen or in the picture, including other people’s content. Both files are stored as JPEG images, and we remove their metadata, such as location, before they are stored. A person who is not signed in can send a problem report too: text and the snapshot above, no files, no link to an account, and the network address is not stored in the report (it is used for a limit of three reports a day, kept as a keyed hash that deletes itself after 24 hours).
• Device and usage data: app version and platform, IP address in server logs, and diagnostics (launch and navigation timings, errors, a random installation identifier; see section 4).

Our infrastructure provider may separately process network metadata to operate its network.

Visitors without an account. A visitor who is not signed in can browse public content only. We do not create an account or a server session for such a visitor and do not store their data in the App beyond local settings such as language and theme. Creating, saving, liking, importing, reporting or any other action that writes data requires signing in. The one exception is sending a problem report or idea, which is protected by a Cloudflare Turnstile check and a per-address limit.

2. How we use information

We use data to:

• create and operate your account and authenticate you,
• store and synchronize lists and wishes, including shared and family lists,
• show your lists and profile to the people you choose, and your public lists in search and Inspiration while you have Plus (section 3, “What other people can see”),
• track prices of products you ask us to track and notify you of changes,
• extract product details from links you provide and, where the feature is available and you allow it, read the pages you add or import with Anthropic’s Claude (section 3),
• check content against our content rules (automated moderation, section 5),
• provide the paid features of Plus and Plus Family and manage subscriptions,
• send notifications you have not turned off,
• prevent fraud, abuse and policy violations,
• measure and improve product quality, reliability and performance,
• comply with legal obligations,
• provide support and troubleshoot problems, including reading and answering the problem reports and ideas you send.

3. How we share information and who processes it

We do not sell personal information. We use the service providers (processors / subprocessors) below. Each acts on our instructions for the purpose stated and under its own terms and privacy documentation.

Our website’s content management system keeps its own content in Cloudflare D1; it holds no App user data.

Before we switch on a provider or a feature that sends your data to a recipient not listed here, we update this table.

AI processing

We use AI in three ways:

• Cloudflare Workers AI, run in our own Cloudflare account and called from our servers: the automated content check (section 5) and the helper features listed under “Other AI on Cloudflare Workers AI” below.
• Apple’s models on your iPhone, which can read a product page for you (below).
• Anthropic’s Claude, only with your permission (next subsection).

Prompts and outputs are not used by us to train models. Do not put sensitive personal details in product or list names used for these features.

Reading pages with Claude (Anthropic), only with your permission

Reading pages with Claude is offered to Plus subscribers in the countries Anthropic serves. It is being switched on step by step: until it is switched on, we send nothing to Anthropic, even if you have already allowed it in the App.

For Plus accounts, the App can send the text of the page of a link you add, or of a page you import, to Anthropic’s Claude (model Claude Haiku) so it can read the product details or the list. Before the first time, the App asks you with a sheet titled “Read links with Claude?”, and sends nothing unless you tap Allow. We record which version of the text you agreed to. If you tap Not now, or if you have not agreed, we do not send the page to Claude; the simple reading of the page still runs, and the wish or import may find fewer details.

• When. When you add a wish from a link (including a Quick Add you move into a wish from its link, and a retry of that read that we queued for you), when you change a wish’s link, and when you import a list from a page. Free accounts and visitors do not use Claude. For an import, Claude reads the page only when the simple reading finds fewer than two items on it.
• What is sent. A bounded piece of text from the page: for a product, about 3,000 tokens of the page, the product part first; for an import, about 6,000 tokens of the page. For a product, the text is what our server read from the public page. For an import, it is what your device read. An import page your device read can show you personal details because you are signed in to the shop, for example your name, address or orders; that text is sent as it is on the page. Avoid importing pages like that while you are signed in to a shop if you do not want them to go to Anthropic.
• What we do not add. We do not add your name, your e-mail address, your account ID, what you typed for the wish, or your other wishes and lists. An imported page may show your name or other details; it is sent as it is (above). We do not send your location; we use your country only to choose the server the request leaves from. With the request goes a coded ID of your account (a keyed hash made with a secret only we hold, so it does not show who you are): it is the same for all your requests. Cloudflare’s documentation does not say whether its AI Gateway passes that coded ID on to Anthropic, so it may be sent.
• Where. The request leaves Cloudflare from a server in the United States, or from a server in the European Union if you are in the EU, the EEA, the United Kingdom or Switzerland, and goes to Anthropic in the United States. We use Claude for you only where you are in a country Anthropic serves: we read your country from your request while it is made and do not store it (a retry we queued for you carries only the country code until it runs). The transfer rests on the Standard Contractual Clauses (with UK and Swiss addenda) in Anthropic’s Data Processing Addendum (read 2026-11-02; the EU-U.S. Data Privacy Framework is not claimed in Anthropic’s documents).
• What Anthropic does with it. Anthropic states that it does not use inputs and outputs of its API to train its models by default, and deletes them within 30 days; if Anthropic flags a request as a violation of its usage policy, it may keep that content longer, up to 2 years. Anthropic does not delete specific data on request for API customers, so we cannot ask it to delete your page text earlier (Anthropic’s published terms, privacy.claude.com articles 7996868, 7996866 and 7996875, read 2026-11-02).
• What we keep. Cloudflare AI Gateway keeps a log entry of every request: time, model, the number of tokens, cost, the coded ID and the kind of request. The page text we sent and the answer are never kept there for a request made for you. The log has no time limit: each of our gateways keeps at most 100,000 entries and deletes the oldest when it is full (read from our Cloudflare account on 2026-11-02). What we store from the answer is the product details written to your wish.
• Turn it off. In the App, open Settings → Privacy and safety and switch off “Read links with Claude”. The same switch turns it on again. After you turn it off we send nothing more to Claude for you, including the retries we had queued. What was already sent stays with Anthropic for the period above; our log keeps only the entry without the page text.
• Reading rules for shops. Once this feature is switched on, our system also uses Claude to keep our reading rules for shops’ public pages up to date: it sends the cleaned text of a public product page that our server itself rendered, with the page’s web address, and the request and answer are kept in the AI Gateway log. No account is behind such a request and no personal data is meant to be on such a page. It does not ask for your permission, because it is not about you or your content; a link you add can lead our system to read that public page.

Safety checks do not depend on this permission. The automated check of the text and photos you post, including the photos you upload (section 5), runs on Cloudflare Workers AI whether or not you allow Claude, because it protects other people. Your permission for Claude does not change it, and the check does not send your content to Anthropic.

Other AI on Cloudflare Workers AI

• Product-page extraction. Deterministic extraction without AI is available to everyone and stays on our servers and your device. The AI step that reads a product page is Claude, described above.
• Checking how we read a shop’s pages. For many shops we read product pages with saved rules that say where the product name and price are on the page. When our server itself reads such a page (not your device), for any account and for the automatic price updates of tracked products, Workers AI may check that the rules found the right product name and price. It receives the page address, the page title and main heading, the name and price the rules found, up to eight other prices on the page with a few words around each, and up to 600 characters of the area around the price. It does not receive what you typed for the wish. The model only confirms or rejects what the rules found; it does not choose other values or write new rules. When the check runs for an account without the AI features of Plus, or for an automatic price update, it is attributed to our system, not to an account.
• Quick Add list suggestion. Only when you tap Suggest a list, we send the Quick Add product name and the names of your eligible personal lists. We do not send the Quick Add URL, note, account ID or list IDs, and the item is never moved automatically.
• Wish categories (Plus). We may send a wish name, description, link address, shop name and the App language to suggest a category for filters; a category you choose yourself is not replaced.
• Price-tracking check. When the price of a tracked product falls, we may send the product name saved for that page and the name found on the page to check that both are the same product.
• Automated moderation (section 5). The names and descriptions you type for a wish or list, your profile name and description, text that a shop page gave a wish, photos you attach to a wish, and the shop’s photos of the product on a wish (our server downloads such a photo for the check and does not keep it). The text of a report you file is not sent to the AI service.
• On your iPhone. On an iPhone with Apple Intelligence the App may read a product page with Apple’s models. Apple runs them on your device, or, where Apple routes a request to its Private Cloud Compute, on Apple’s servers, which Apple states do not keep the data.
• Hybrid semantic search is off by default. If we enable it, Workers AI receives your search query and bounded titles, tags and descriptions from indexed items and profiles to create embeddings. This can include private wishlist content so it can be found in your own search; account IDs and resource IDs are not sent. If gated reranking is enabled, it receives the query and bounded text of public search candidates only, and only reorders results.

Account IDs, resource IDs and Quick Add URLs are not sent with these requests. Workers AI requests are not logged with their content.

Importing a wishlist from a link

When you import a wishlist or product page from a link, the App opens the page on your device and may send the page’s HTML, as rendered on your device, to our server so we can read the structured product data it contains. The HTML may include whatever the website shows to you, for example a personalized or logged-in page. We use it only to extract the list and its items, do not store it with your account, and do not cache it. If no device HTML is provided, our server fetches and renders the page itself using Cloudflare Browser Run. Importing is limited to public pages. The AI step of an import is Claude, available to Plus subscribers with your permission once the feature is switched on (section 3, “Reading pages with Claude”): if the simple reading of the page finds fewer than two items on it, the text of the page your device rendered, about 6,000 tokens, may then go to Anthropic. That text is not kept in our AI logs. If you have not allowed it, the import asks first and does not send the page otherwise.

Photos from shops and other websites

A wish may show a photo of the product from a shop’s website, from an image host the shop uses, or from another website whose photo link you paste. We do not store copies of these photos. The App loads them directly from that website whenever the wish is shown, so the website receives the request from the viewer’s device, including the device’s IP address and the usual technical details of the request, and handles them under its own privacy policy. This applies to everyone who views the wish, including visitors of a public list.

To check such a photo (section 5), our server downloads it, checks it and does not keep the downloaded copy; we keep only the result of the check, keyed by the photo’s address (section 7). The website sees this download as a request from our infrastructure provider, not from you. If we must preserve a wish for a legal reason (for example a report to child-safety authorities), we may download its photo again and keep that copy in private storage for as long as the law requires (section 7).

Photos you upload from your phone are stored in our storage (Cloudflare R2), linked to your account, and deleted with the wish or with your account.

Shopping links and recommendations

We do not use affiliate links or show shop recommendations. If we start, we will update this policy first, naming the provider and the data it receives.

What other people can see

Your lists and your profile are private until you choose otherwise.

• Private lists. Only you can see a private list, together with the people you share it with (in the role you give them) and, for a list of your family group, the members of the family.
• Share links. Anyone who holds a share link sees a short preview before joining: the list name, your name and profile photo, the role offered and when the link expires. You can revoke a link at any time.
• Public lists. Making a list public is part of Plus. While you have Plus, anyone who has the link can open your public list, including people who are not signed in, and it is shown on your profile page to people who can open your profile. If your Plus ends, your public lists stay marked public but nobody else can open them until you have Plus again.
• Search and Inspiration. While you have Plus, your public lists and their wishes can also be found in search, in Inspiration and in popular tags.
  • On the “Popular lists” shelf a list is shown with its name, the number of its wishes and up to three photos of them. Your name, username and profile photo are shown with it only if your profile is public.
  • The “Stores” shelf counts the wishes of such lists by shop (for example, ikea.com) and shows photos of popular wishes from that shop. It does not name anybody.
  • If your Plus ends, your public lists leave search and Inspiration.
  • Wishes and lists rated 18+ never appear on the Inspiration shelves. In search, only signed-in people who turned on “Show 18+” and chose to see 18+ results can find them.
• Your profile. Your profile (name, username, description and profile photo) is private by default. If you make it public, anyone with its link can open it and see your public lists, and people can find you in people search. While it is private, only you and people who can see one of your lists through a share or your family group can open it, and you do not appear in people search. A private profile does not hide a list you made public: the list stays open by its link and can be found in search and Inspiration as described above. To keep a list away from people you do not know, make the list itself private.
• Followers. A person can follow you only if they can open your profile. Anyone who can open your profile sees how many people follow you and whom you follow, and who they are. While they can open your profile, your followers see your public activity in their feed: lists you make public, wishes you add to public lists, public wishes you like, people you follow, and wishes on your public lists that you mark as received. If they can no longer open your profile (for example, after you make it private), they stop seeing it. Feed records are kept for 365 days (section 7).
• Reservations and purchases. People who share a list see which of its wishes are reserved or bought, so nobody buys the same gift twice. What the owner of the list learns depends on the list’s setting: nothing (the default), only that a wish is taken, or who took it, if that person agreed to show their name. Visitors of a public list see no marks.
• Family group. Members of your family group see the group, its members and their roles, and the lists of the group.

Other people see new or changed content only after our automated check has approved it (section 5). This applies to every list, private, shared or public, and to your profile.

4. Analytics and diagnostics

• Diagnostics. EAS Observe collects app launch and navigation timings (screen names with ids filtered out), device state such as battery and connection type, a random installation identifier, and error reports. We configure it not to send names, e-mail addresses, wish text, URLs, tokens or user identifiers, and we send it only fixed event names. The text and trace of an unexpected error are sent as they are and can contain personal data by accident; EAS offers no switch for that other than turning error reports off. The installation identifier is also included in a problem report you send, so for those reports our diagnostics data is linked to your account.
• Usage metrics. Our servers record usage data points in Cloudflare Analytics Engine to keep the service running and within budget: AI usage (the feature, your plan tier, the outcome and the amount used, under a coded ID of your account, a keyed hash that does not show who you are), how we read shop pages (the shop’s domain and the outcome), and the health of our queues and notifications. These data points do not contain the text of your content.
• No product analytics events. We do not currently record a log of what you do in the App for product analytics. If we start, we will describe it here first.
• No advertising. The App contains no advertising SDKs. We do not use advertising identifiers, we do not share data with advertising networks, and we do not track you across other companies’ apps or websites.

5. Automated content moderation

Content that others can see is checked automatically. No human reviews it in the standard process. A person decides an appeal of a photo that no automatic check can judge, an appeal of a takedown made by a person, a report, a notice of illegal content and a legal hold.

• What is checked. Every new or changed wish, list and profile text, and every photo on a wish, is checked by a Cloudflare Workers AI model (Cloudflare “Clef”), which returns approved or rejected. Saving never waits for the check: it is queued when you save.
• Who sees what before the check. You see your own wishes, lists and photos at once. Everybody else, including the people you share a list with, your family group, visitors of a public list or profile, search, feeds and discovery, sees new or changed content only after the check has approved it. The co-authors of a list (the people who can edit it) see that a change is waiting for the check.
• Photos from shops and other websites. Our server downloads such a photo for the check and does not keep it. If the same image address was checked before, we reuse that result for your wish, at no new check; otherwise one check is made and its result serves every later wish with that address. If the automatic check rejects an address, the photo is rejected in every wish that holds it, in every list, including wishes where an earlier result had approved it; the owner of each of those wishes gets the usual in-app notice with the reason category and may appeal. A copy of a rejected photo added later takes the same rejection with no new check, and the way back is an appeal; only a person (an administrator who upholds an appeal) clears a rejected address, and a later approval by the automatic check does not. This happens only when the automatic check itself rejects the photo, not for a legal hold or a decision of a person. A photo from another website stays at its address and we do not copy it, so the website can change what it shows there after our check; a report makes us look at the photo as it is then. The link to the photo or to the shop page is chosen by the person who adds the wish, so a photo is never trusted for coming from a shop.
• Photos we cannot check. Only a photo that is really gone is removed from the wish: the website says it does not exist (404 or 410), or its address is one our safety guard refuses. A photo we cannot read for any other reason (the website refuses our server with 401 or 403, a security page or a page or an empty answer instead of an image, a format we cannot read, a photo larger than 3.75 MiB, or timeouts that go on) is kept for you, marked “Photo not checked”, and hidden from other people. We do not check it again by ourselves: it stays hidden from other people until you replace it with another photo.
• If the check cannot run (for example because a safety budget is used up), the content stays hidden from other people until it has been checked.
• Rejections and appeals. If content is rejected, it remains visible only to you with the reason category, and you may edit it and resubmit, or appeal once. An appeal triggers one more automated check with a stronger model and a different threshold. An appeal of a photo too large for the model is decided by a person instead. If you disagree with the result of an appeal, write to support@sparkhop.app and a person will look at it.
• Reports. A report you file hides the content for you at once. The AI service does not read the text of a report; our administrators read reports and decide them. Blocking a user always takes effect immediately.

We may later stop checking in advance the text of small private lists that only the people you invited can open. We will update this policy before such a change.

18+ content and age confirmation. The automated check also rates approved content as general or 18+ (see the Content and Moderation Policy). 18+ content is hidden by default and is shown to other people only if they turn on “Show 18+” in their settings; it is never shown to visitors who are not signed in. Turning the setting on requires confirming that you are at least 18 and a separate consent. For this we store, per account, only: whether the setting is on, that an age of at least 18 was confirmed, the source of the confirmation (an age signal from your device or app store, or a declared birth date) and the time of confirmation. If the age signal of the app store (on Android, Google Play Age Signals) is used, we receive only the lower bound of the age range, not your birth date. If you enter a birth date on a neutral age screen, we use it once to work out whether you are at least 18 and do not store it. If an age under 18 is reported or entered, we store the fact that the setting is locked for the account (with the time) and do not offer it again. On iOS the setting is turned on on our website, not in the app.

See also the Content and Moderation Policy.

6. Legal bases

Where applicable law requires a legal basis, we process data on the basis of:

• contract performance (accounts, lists, subscriptions, notifications you request, the record of the Terms version you accepted, and showing the lists you share or make public, including in search and Inspiration with Plus),
• legitimate interests (security, abuse prevention, moderation, reliability, including checking with AI that we read shop pages correctly, diagnostics, and reading problem reports to fix problems),
• consent (push notifications, optional photo and camera access, and sending pages to Anthropic’s Claude; you can withdraw each in the App or in your device settings),
• legal obligations (tax and accounting, responding to authorities).

Where we rely on legitimate interests, you may object (section 11).

7. Retention

We keep data only as long as needed for the purposes above. Specific periods:

Some records may be kept longer where required for legal compliance, fraud prevention, accounting, dispute resolution or security.

8. Cookies and similar technologies

The App stores local settings and a session on your device. Our website sparkhop.app sets no cookies and runs no analytics or advertising scripts; it remembers the language you chose in your browser’s local storage. Where you sign in on the web, we use a cookie that keeps you signed in, which is needed for the service. Our Turnstile checks (Cloudflare) may run on sign-in and on sensitive actions to detect automated abuse.

9. International transfers

Our providers operate globally, so your data may be processed in countries other than your own, including outside the European Economic Area and the United Kingdom. The production database is hosted by Turso in the USA (AWS region us-east-1). Our code runs on Cloudflare’s network, which places it automatically, and Cloudflare chooses where our file storage is kept; we do not fix a country for either. Where the law requires it, we rely on the safeguards our providers offer in their data processing terms, such as the Standard Contractual Clauses approved by the European Commission and their UK addendum.

Pages sent to Anthropic (Claude). If you allow it and the feature is switched on (section 3), the page text described there goes to Anthropic PBC in the United States, from a Cloudflare server in the United States or, for people in the EU, the EEA, the United Kingdom and Switzerland, in the European Union. We do this only where Anthropic serves your country. The transfer rests on the Standard Contractual Clauses (with UK and Swiss addenda) in Anthropic’s Data Processing Addendum; the EU-U.S. Data Privacy Framework is not claimed in Anthropic’s documents.

People in Japan. The page text goes to Anthropic PBC in the United States only because you allowed it. What is sent: the text of the page you add or import. Who receives it: Anthropic PBC. Why: to read the product details or the list.

People in Korea. What: the text of the page you add or import. Country: the United States. When and how: at the time you add or import, over the internet. Recipient: Anthropic PBC. Purpose: to read the product details or the list. Period: Anthropic deletes it within 30 days, or keeps it up to 2 years if it flags the request as a usage-policy violation, and does not delete specific data on request (Anthropic’s published terms, read 2026-11-02). You can refuse in the App, Settings → Privacy and safety; the rest of the App keeps working.

10. Children

You must be at least 13 years old to use Sparkhop, or older where the law of your country sets a higher age for agreeing to the processing of your personal data. The App is not directed to children below that age. If you believe a child has given us personal data, write to privacy@sparkhop.app and we will delete it.

11. Your rights

Depending on where you live, you may have the right to access, correct, delete and export your data, to restrict or object to its processing, to withdraw consent at any time (this does not affect processing before the withdrawal), and to lodge a complaint with a supervisory authority, for example the one in the country where you live or work.

• In the App. You can delete your account (see our Account Deletion page), download a copy of your own data (account, lists, wishes, settings and more) as one file with “Download my data” in Settings, and turn off sending pages to Anthropic’s Claude in Settings → Privacy and safety.
• By e-mail. For anything else, write to privacy@sparkhop.app. We may ask you to confirm that the account is yours, usually by writing from the e-mail address of the account. We answer within one month; where a request is complex, we may extend this by up to two more months and will tell you why.
• United States. We do not sell personal information and do not share it for cross-context behavioural advertising. Residents of any US state can use the rights above in the same way, and we will not treat you differently for using them.

12. Account deletion

You can delete your account in the App (Settings → Delete account) or by writing to privacy@sparkhop.app. Our Account Deletion page explains the steps. Deletion removes or anonymizes:

• your profile, sessions and sign-in links,
• the record of the Terms version you accepted,
• your lists, wishes, photos, Quick Adds, likes, shares and invitations,
• your problem reports and ideas, with their screenshots and photos,
• your notifications, notification settings and push tokens,
• your reservations on other people’s lists; a wish you marked as bought stays marked, without your name, so nobody buys it twice,
• your family membership (a family in which you are the only member is deleted with you; if you were its only owner, the member who joined first becomes the owner),
• your cached data and per-user processing state at our infrastructure providers.

Deleting your account does not cancel a subscription: cancel it in your App Store or Google Play subscription settings. Subscription and billing records that the store requires for accounting may be kept, de-linked from your account. Some records may be kept where required for legal compliance, fraud prevention, accounting, dispute resolution, or security logs: for example reports other people filed about you, notices of illegal content, information we must keep by law (section 7) and subscription records for the periods in section 7.

13. Changes to this policy

We may update this policy from time to time and will update the effective date when we do. For material changes we will notify you in the App before they take effect.

14. Contact

Questions about this policy and privacy requests: privacy@sparkhop.app. Other questions: support@sparkhop.app. We answer in English or Russian.