Arkanika

Arkanika · Legal

Arkanika Privacy Policy

Effective date: 5 August 2026

This Privacy Policy explains how we handle personal information in Arkanika, a dark-mystical, AI-powered tarot app for iOS and Android. It applies to the Arkanika mobile app and any companion web pages we operate (including our web deletion page at arkanika.net/delete-account), everywhere the app is available.

Arkanika is strictly for adults aged 18 and over. It is not directed to children, and we do not knowingly collect data from anyone under 18.

We may update this policy from time to time (see "Changes to this policy" below).


The short version (what you most need to know)

  • Your readings are written by AI. When you create a reading or chat with your mage, the details you enter — plus your display name and a short summary of your journey in the app (your streak, and the themes and cards of your recent readings) — are sent to our AI partner Anthropic to write your reading. Anthropic works only on our behalf under contract, never uses your data to train its AI, and automatically deletes it within 30 days (kept only for safety monitoring).
  • We never sell or share your reading or chat content, and we never use your readings or chats for advertising.
  • We do not sell your reading data. If you see ads (you have neither Premium nor the Remove Ads purchase), your advertising identifier may be "sold"/"shared" for personalized ads under California law — you can opt out at any time, and personalized ads in the EEA/UK happen only with your consent.
  • You're in control. You can withdraw your AI permission, and export or permanently delete everything, from Settings in the app (or delete via arkanika.net/delete-account without opening the app).

This summary is only a guide. The full policy below controls.


1. Who we are

Yurii Onyshko, an individual entrepreneur (фізична особа-підприємець, "ФОП") registered under the laws of Ukraine, with registered address 14A M. Khvylovoho St., Lviv, 79000, Ukraine, is the data controller (under the EU/UK GDPR) and the business (under the California CCPA/CPRA) responsible for the personal information described in this policy.

Data Protection Officer: We have not appointed a Data Protection Officer because we are not required to. You can reach our privacy team using the contact details above.

2. Contact for EEA and UK users

For individuals in the European Economic Area and the United Kingdom, the direct contact for all data-protection questions and requests is [email protected]. We respond to all data-protection matters through that address, wherever you are.

3. Age: Arkanika is for adults 18+

Arkanika is strictly for adults 18 and over. It is not directed to children, and we do not knowingly collect personal information from anyone under 18. When you start using the app, you self-attest that you are at least 18, and we record that attestation. Our store age ratings reflect an adults-only (18+) filing.

If we learn that someone under 18 has used the app or provided us data, we will delete that account and its data. Because the app is for adults only, a parent or guardian cannot consent to a minor's use on their behalf. If you believe a minor has provided us information, contact us at [email protected].

4. Data we collect, why, and our legal basis

We collect only what we need to run Arkanika. The table below lists every category of data, why we use it, and — for EEA/UK users — our legal basis under Article 6 of the GDPR. Retention is summarized in section 8.

Legal-basis key: Contract = necessary to provide the service you asked for · Consent = your freely-given, withdrawable consent · Legitimate interests = our legitimate business interests, balanced against your rights · Legal obligation = required by law.

Identity & account

Data Why we use it Legal basis
Anonymous guest identifier (auto-created on first launch — no login needed) Create your account, own your data, run the game economy Contract; Legitimate interests (basic abuse/fraud prevention)
Sign in with Apple / Google (optional): provider user ID + email (Apple may give a private "relay" email) + the name your provider shares at sign-in (if any) Keep your account across devices; re-sign-in; address you in account emails Contract; Consent (your choice to link)
User-chosen display name (not your real store-account name) Personalize greetings, the UI, and your readings & chat (sent to our AI partner as part of writing them — see section 6) Contract
Timezone Calculate your daily reset, streaks and moon timing Contract
18+ age-attestation flag Enforce our adults-only rule and comply with app-store age policies Contract; Legitimate interests (store-policy compliance, age-gating)
AI-permission timestamp Record the in-app permission you gave before any AI reading is generated, and keep proof of it Legitimate interests (operating the AI permission gate); Legal obligation (GDPR Art. 7(1) requires us to keep records of consent)

Reading content — sensitive, written by you

Data Why we use it Legal basis
The question you type, who the reading is "for", the theme you choose Generate a personalized reading (we send these to our AI partner to write it) Contract — necessary to deliver the reading you requested. Where your free text may reveal special-category data, we rely on your explicit consent under GDPR Art. 9(2)(a) (see section 5).
The cards drawn Generate the reading; track collection progress Contract
The AI-generated reading text and your saved reading history Deliver and let you revisit your readings Contract
Personalization details we already hold: your display name, your current streak, and a short summary of your recent readings (their themes, and cards that have recurred) Make your reading and chat replies feel personally yours (sent to our AI partner with the request — see section 6) Contract

Why this is sensitive: what you type is free text and can reveal things like your health, relationships, beliefs or finances — even another person's information. We never ask for this, but it can appear. See section 5.

Chat with your mage — sensitive

Data Why we use it Legal basis
Your messages to the mage (plus earlier chat turns, the related reading content, and the personalization details above, re-sent so the conversation stays coherent and personal) Continue the conversation about a reading Contract — necessary to deliver the chat you requested; explicit consent (Art. 9(2)(a)) covers any special-category data in your free text
The AI-generated mage replies Deliver and keep the thread Contract

Progression & game economy

Data Why we use it Legal basis
XP, level, energy, streak, moon charms, ritual state, 78-card collection, unlocks Run the gamified experience Contract

We do not share your economy state (your balances) with third parties for its own sake — it lives in our database. One narrow exception: your current streak is included in the personalization details sent to our AI partner when writing a reading or chat reply (see section 6).

Purchases & subscriptions

Data Why we use it Legal basis
Subscription/entitlement status and purchase history (via RevenueCat: app user ID + purchase/receipt info) Manage Premium, remove ads, restore purchases Contract; Legal obligation (retaining financial records for tax/audit under applicable tax law)
Payment card details We never see these. Apple and Google handle all payment.

Content you report to us (safety)

Data Why we use it Legal basis
Reports you submit using the in-app Report control: the reason you select, any detail you add, and a reference to the flagged reading or conversation, linked to your account Review reported content (a person on our team may read it), keep the app safe, act on abuse, and improve our safety guardrails Legitimate interests (safety, abuse prevention, improving guardrails); Legal obligation (responding to lawful requests and notice-and-action duties, e.g. under the EU Digital Services Act)

Note: Readings and mage replies are AI-generated or drawn from content prepared in advance, and are not individually reviewed by a person before you see them. The exception is content you choose to report — a person on our team may then review the referenced content. See section 17 of this policy and Section 12 of our Terms of Use.

Support & suggestions

Data Why we use it Legal basis
Support & suggestion messages you send us via the in-app form (the category you choose + your message text, linked to your account) Answer your request, act on your problem report or suggestion, improve the app Contract; Legitimate interests (providing support)

Launch waitlist (website)

Data Why we use it Legal basis
Waitlist email address you submit on arkanika.net (plus its confirmation state and, if you arrived through a friend's referral link, the referral code that brought you — a code, not a person's identity) Hold your place in the launch queue, send you the confirmation email you must click (double opt-in) and, at launch, one "we're live" email with your reward code; count confirmed referrals toward the advertised reward (capped) Consent — you join by submitting the form and confirm by clicking the emailed link; leave anytime by writing to [email protected]

The waitlist lives outside your app account: it is never linked to your readings, chat, or game data, and deleting one never touches the other. Waitlist emails are used only for the two emails above — never marketing beyond them, never shared.

Device & technical

Data Why we use it Legal basis
Device identifiers, OS/device model, app version Compatibility, support, security Legitimate interests; Contract
Push notification token (a device token issued via Expo, created only if you turn notifications on) Send the optional reminders you asked for: a daily-card nudge and a streak warning Consent — you enable notifications through an in-app choice plus the operating system's permission prompt, and can turn them off at any time (in the app or in your device settings)
IP address — including a keyed, hashed daily counter of signups per IP used to limit abuse (for the counter we store only the hash, never the raw address; kept 7 days) Security, abuse/fraud prevention, routing Legitimate interests (security and fraud prevention)
Crash & error diagnostics — via Sentry (PII-scrubbed, EU data region) and via error reports the app sends to our own server (error message, technical stack trace, app version and platform, linked to your account) Diagnose crashes and errors, keep the app stable Legitimate interests
Device advertising identifier (Apple IDFA on iOS / Google Advertising ID on Android; AdMob — users who see ads only, i.e. neither Premium nor the Remove Ads purchase) Serve ads; personalize ads only with consent Consent (EEA/UK — ads are shown only after Google's consent flow, whether personalized or not); Legitimate interests (outside the EEA/UK, where consent is not legally required)
Internal AI usage/cost metrics (token/cost metadata only — no reading or chat content) Monitor our AI costs and reliability, and enforce per-user daily limits Legitimate interests (cost and reliability monitoring). These metrics are linked to your account while it exists and are deleted together with your account (see section 8); they are included in your data export.

Where we rely on legitimate interests, we have weighed our interest (for example, keeping the service secure and improving it) against your rights and freedoms, and we use the data in ways you would reasonably expect. You can object — see section 10.

5. Sensitive information

The question you type, who a reading is for, the theme, and your chat messages are free text. They can reveal sensitive things about you — health, sex life, relationships, religious or philosophical beliefs, finances — or information about other people, even though we never ask for it.

Please don't enter anything you wouldn't want processed, and avoid entering other people's personal or sensitive details.

  • How we generate your reading (GDPR): The transfer of your reading and chat inputs to our AI partner is necessary to provide the reading or chat you requested, so our lawful basis for that transfer is contract (Art. 6(1)(b)). Separately, because your free text may reveal special-category data (Art. 9), we ask for your explicit consent under Art. 9(2)(a) as the condition for processing any such data, and — to meet Apple's App Store requirements — we also ask for your explicit in-app permission before any reading is generated. Our server will not generate a reading without that in-app permission on file. You can withdraw it at any time (see section 13); if you do, we can no longer generate readings or chat, because that processing is essential to those features.
  • California (CPRA): this content may be "sensitive personal information." We use it only to provide the reading and chat you asked for — never to infer characteristics about you and never for advertising. Because we already limit our use to what's necessary to deliver the service, there is nothing further to limit, but you can still withdraw your AI permission and delete this content at any time.

We never use your reading or chat content for advertising, and we never sell or share it.

6. Who we share data with (at a glance)

We share personal information only with the service providers and partners below. Most act as our processors/service providers — they act only on our instructions under a data protection agreement (DPA). A few are independent controllers for a specific relationship (payments, or ad personalization).

Recipient Role What it receives Notes
Anthropic, PBC Processor / sub-processor Reading inputs (your question, who it's for, theme, cards), your chat messages, and personalization details (display name, current streak, recent-reading summary) AI generation of readings & chat. Under a data processing agreement; Anthropic does not train on your data and automatically deletes inputs/outputs within 30 days (kept only for safety monitoring). This is our headline third-party-AI disclosure.
Railway Corp. Processor Effectively all stored data Cloud hosting of our API server and database (region: European Union — Amsterdam, the Netherlands)
RevenueCat Processor App user ID + purchase/receipt info Subscription & entitlement management
Apple / Google Independent controllers Sign-in identity + store billing/payment Also our sign-in providers; we never see your card details
Google AdMob Independent controller (ad personalization) Advertising identifier, IP, ad-request signals — users who see ads only (neither Premium nor the Remove Ads purchase) Consent required in EEA/UK (via Google UMP) and via Apple ATT on iOS; may be a "sale" and/or "share" under CPRA (see section 11)
Sentry Processor Crash/error diagnostics (PII-scrubbed) Stability & error diagnosis
Expo Processor Basic device/app-version info when the app checks for updates; if you enable notifications: your device push token and the content of the notifications we send (short, generic reminder lines — never your reading or chat content) App runtime, over-the-air update delivery & push-notification delivery
Resend, Inc. Processor Your email address + the content of account emails we send you (e.g. deletion confirmations) Transactional email delivery (EU sending region), under a DPA
GitHub, Inc. (Microsoft) Processor Private nightly backup copies of our database (encrypted at rest by GitHub) Backup storage in the United States; each backup is automatically deleted after 30 days
Cloudflare, Inc. Processor Website visits to arkanika.net (IP address, request metadata) and the routing envelope of email sent to our @arkanika.net addresses Hosts, proxies and secures our website (including the web deletion page) and routes inbound email to our mailbox (US company)
Google (as our mailbox provider) Processor Emails you send to support@ / [email protected] (your address + your message) Our support/privacy mailbox is hosted with Google; mail to our addresses is delivered and stored there

Our processors are bound by DPAs and may act only on our instructions. We keep our list of sub-processors current and can provide details on request at [email protected].

We never share your reading or chat content for advertising, and we never sell it.

More on the AI (Anthropic)

To write your reading and generate mage replies, we send Anthropic the details you provide — your question, who the reading is for, the theme, and the cards drawn — together with personalization details we already hold: your chosen display name, your current streak, and a short summary of your recent readings (their themes, and cards that have recurred), so the writing can feel personally yours. During a chat, we also re-send your chat messages, the earlier turns in that conversation, and the related reading content (the question/theme/cards and the reading text) so the multi-turn conversation stays coherent. We do not send Anthropic your account identity beyond the request itself, your email, the name your sign-in provider shares, payment data, advertising ID, or your game-economy balances beyond the streak just described.

  • Legal basis: the transfer is necessary to deliver the reading or chat you requested (contract), and your explicit consent covers any special-category data in your free text. We also require your explicit in-app permission (recorded with a timestamp) before making any AI call; without it, our server returns an error and makes no AI call. You can withdraw that permission in Settings.
  • Safeguards: a data processing agreement with Anthropic (incorporated into its commercial terms). Anthropic does not use your inputs or outputs to train its models and automatically deletes them within 30 days (kept only for safety monitoring). The lasting copy of your reading lives in our database and is erased when you delete your account.

This matches what the in-app consent screen tells you, word for word:

"Your spread readings and mage chats are written for you by an AI oracle. When you ask for a reading — and when you chat with your mage — what you share (your chosen theme, who it's for, your messages) is sent to our AI partner (Anthropic) to write the reply, together with your display name and a glimpse of your journey (your streak, and the cards and themes of recent readings) so it feels truly yours. What you write is up to you — it may reveal tender things (your health, your beliefs, your heart); by agreeing you explicitly consent to us handling whatever you choose to share, solely to create your readings and chat. The daily card is served from writings prepared in advance — nothing personal is sent for it. We never sell your readings or chats. You can withdraw consent, and delete everything, anytime in Settings."

7. International data transfers

We and some of our providers process data outside your country, including in the United States and in the European Union — Amsterdam, the Netherlands (where our database is hosted). US-based recipients include Anthropic (our AI partner), RevenueCat, Sentry (a US company; our crash data is stored in its EU region), Resend (a US company; EU sending region), Expo (app updates and push delivery), GitHub (backup storage), Cloudflare (website hosting and email routing), and Google (AdMob, sign-in, and our support mailbox).

Because we are based in Ukraine, we also access and administer data from Ukraine when operating the service. For that access we remain directly bound by the GDPR (Article 3(2)) and apply the safeguards described in section 9; the European Commission is preparing standard contractual clauses for controllers outside the EEA that are directly subject to the GDPR in this way, and we will adopt them for this scenario when they are finalised.

Where we transfer personal data out of the EEA or UK, we rely on appropriate safeguards, including the EU Standard Contractual Clauses and the UK International Data Transfer Addendum, together with supplementary measures. For recipients that act as independent controllers (for example, Google AdMob for ad personalization), the recipient's own transfer mechanism applies to the data it controls. You can request a copy of the safeguards we use by emailing [email protected].

8. How long we keep data

Data How long we keep it
Guest & account data (including readings, chat, economy) Until you delete it — except abandoned guest accounts, below
Abandoned guest accounts (never connected to Apple/Google, not premium) Deleted after about 6 months with no activity in the app
Account deletion Two-tap in-app → soft-delete, then a 30-day hard purge that permanently erases your data; if you linked Apple sign-in, we call Apple's token-revocation on deletion
Content you report to us (safety reports) Kept as long as needed for safety review and record-keeping, then deleted; retained longer only where a legal obligation requires it
Short-lived logs Held briefly by our hosting provider as part of running the service (on the order of 30 days)
Anthropic (AI partner) Automatically deleted within 30 days (never used for training)
Purchase/financial records Remain with the app store (Apple/Google) and our subscription processor as required for tax/audit; once your account is deleted they are no longer linked to your Arkanika data
Internal AI usage/cost metrics Linked to your account while it exists; deleted together with your account
Crash data (Sentry) Retained by Sentry per its standard retention (currently 90 days), then deleted
Self-hosted app error reports Until your account is deleted (removed with the account)
Push notification token (if you enabled notifications) Until your account is deleted (removed with the account); notifications stop immediately when you turn them off in the app or revoke the OS permission
Support & suggestion messages Until your account is deleted (removed with the account)
Hashed signup-IP counters 7 days
Web deletion-request records (the email you submit + any optional note) 180 days, kept as an audit trail of your request
Launch waitlist emails (website) Unconfirmed: erased after 30 days. Confirmed: kept until the launch mailing is sent and the waitlist winds down, or until you ask us to remove you ([email protected]) — whichever comes first
Database backups Rolling nightly copies, each automatically deleted after 30 days

Where no fixed period applies, we set retention based on why we need the data, legal requirements, and the risk involved.

9. How we protect your data

We use technical and organizational measures appropriate to the risk, including: encryption in transit; server-authoritative access controls with per-user authorization checks on every request (each account can only reach its own rows); keeping the AI key server-side only (never on your device); least-privilege access; scrubbing personal data from diagnostics; and a breach-response process. No method of transmission or storage is ever 100% secure, but we work to protect your information and to keep improving our safeguards.

10. Your choices & rights

You have rights over your personal information. To exercise any of them, use the in-app Settings (which offer permission controls, data export, and Delete Account & Data) or email [email protected]. Because most accounts are anonymous, we may verify a request by reference to your account/device or sign-in; we may ask for limited information to confirm you are the account holder. We will not discriminate or retaliate against you for exercising your rights.

If you're in the EEA or UK (GDPR)

You have the right to: - Access the personal data we hold about you; - Rectify inaccurate or incomplete data; - Erase your data ("right to be forgotten"); - Restrict processing in certain cases; - Portability / export — receive your data in a portable format; - Object to processing based on our legitimate interests; - Withdraw consent at any time (as easily as you gave it) — in particular for your AI permission, push notifications, and personalized ads; withdrawal doesn't affect processing that already happened; - Complain to a supervisory authority — your local data protection authority (or the UK ICO).

We do not use your data for solely-automated decisions that produce legal or similarly significant effects. Readings are AI-generated entertainment, not profiling that affects your legal rights.

We aim to respond within the timeframes the law requires (generally one month under the GDPR).

If you're in California (CCPA/CPRA)

You have the right to: - Know / access the personal information we've collected and how we use and disclose it; - Delete your personal information; - Correct inaccurate personal information; - Opt out of "sale"/"sharing" of personal information; - Limit the use of sensitive personal information (we already limit it to providing the service); - Non-discrimination for exercising your rights.

We do not sell your reading or chat data, and we never share it for advertising. As explained in section 11, showing personalized ads to free users via AdMob involves disclosing your advertising identifier and related signals and may be both a "sale" and a "share" under the CPRA, so we provide an opt-out. You may submit requests via [email protected] or in-app Settings, and you may use an authorized agent. We generally respond within 45 days and may verify your request.

California data categories (CPRA disclosure)

The following table maps the statutory categories of personal information to what we do. "Sold/Shared" refers to CPRA's broad definitions, which can include disclosing an advertising identifier for cross-context behavioral advertising.

CPRA category Collected? Sources Business/commercial purpose Disclosed for a business purpose to (categories of recipients) "Sold"/"Shared"? Retention
Identifiers (guest ID, provider user ID, email, name shared by your sign-in provider, device IDs, push token if you enable notifications, IP, advertising identifier) Yes You; your device/app; our providers Provide the service, security, optional notifications, ads to users who see ads Hosting, subscription, crash-reporting, email, backup, notification-delivery, and advertising providers Advertising ID: potentially sold/shared for personalized ads (opt-out available). All others: no. Until deletion; logs ~30 days
Commercial information (subscription/purchase history) Yes Stores; RevenueCat Manage Premium; tax/audit Subscription-management provider No Held by the store and our subscription processor for tax/audit
Internet/network activity (app usage and diagnostic events) Yes Your device/app Stability, security Crash-reporting provider and our hosting provider No See section 8
Geolocation (coarse, inferred from IP; timezone you set) Yes (coarse only) Your device/network Security, routing, daily/streak timing Hosting/security providers No Logs ~30 days
Sensitive personal information (reading questions/themes, chat that may reveal sensitive matters; account log-in via sign-in) Yes You Provide readings/chat only AI partner (to generate content), hosting No — never sold/shared, never used for advertising Until deletion
Inferences No — we do not build profiles or infer characteristics about you No
Other user-authored content (readings, chat, saved history, support messages) Yes You / AI-generated Provide and store your readings AI partner (≤30-day retention), hosting No Until deletion → 30-day purge

11. Advertising, tracking & your "Do Not Sell or Share" choice

Arkanika is free to use with ads; you can remove ads with a Premium subscription or the one-time Remove Ads purchase.

  • Users with neither Premium nor the Remove Ads purchase see ads served by Google AdMob, which uses your device advertising identifier (Apple IDFA on iOS / Google Advertising ID on Android) along with your IP address and ad-request signals.
  • Personalized ads involve cross-context behavioral advertising. Under California's CPRA, disclosing your advertising identifier and related signals to AdMob for this purpose may be treated as both a "sale" and a "share." You can opt out of it at any time: use the Ad privacy options control in the app (Account settings — it opens Google's ad-privacy form where it is available for your region), decline or change App Tracking Transparency for Arkanika in your iOS settings, reset or limit your advertising identifier at the OS level, or email [email protected] and we will complete the opt-out with you (we will confirm your ATT and ad-privacy settings with you and treat your request as a standing opt-out). This opt-out covers both "sale" and "share."
  • In the EEA/UK, personalized ads and any tracking happen only with your consent, collected through Google's consent prompt (UMP) and, on iOS, Apple's App Tracking Transparency (ATT) prompt.
  • Premium and the one-time Remove Ads purchase both remove ads.
  • We never use your reading or chat content for ads, and we never sell or share it.

12. Identifiers & on-device storage

Arkanika is an app rather than a website, but it uses SDK identifiers and on-device storage similar to cookies: - Device advertising identifier (Apple IDFA on iOS / Google Advertising ID on Android; AdMob) — for ads to users who see ads; consent-gated for personalization; - Device identifiers — compatibility, support, security; - Push notification token (only if you turn notifications on) — delivering the reminders you asked for; - Secure on-device storage for your guest/session token.

In the EEA/UK, we treat non-essential SDK identifiers/local storage as consent-based, and only strictly-necessary storage (such as your guest/session token) operates without consent. You can reset your advertising identifier at the OS level, change tracking choices via ATT/UMP, and manage ad privacy in the app via Account → Ad privacy options.

Our website (arkanika.net) stores only your theme choice in your browser and sets no tracking cookies. Our web pages do not currently read browser opt-out signals such as Global Privacy Control — the website neither sells nor shares personal information, so there is nothing there for such a signal to switch off; the in-app and OS-level controls above are the effective ad-privacy controls. If that ever changes, we will honour GPC on the pages it applies to.

13. Consent & how to withdraw it

We rely on your permission/consent for: the AI processing of your reading and chat inputs (your in-app AI permission, plus explicit consent for any special-category data), push notifications, and personalized ads/tracking.

  • We collect your AI permission during onboarding as a separate, affirmative step and record it with a timestamp; we show ATT/UMP prompts for tracking/ads.
  • Push notifications are off by default. They start only after you say yes twice — once on an in-app explanation screen, and once on the operating system's own permission prompt. Turn them off any time in the app (Settings → Moon whispers) or in your device's notification settings; either stops them.
  • Your permissions are granular and can be withdrawn at any time in Settings, without affecting processing that already lawfully took place.
  • If you withdraw your AI permission, we can no longer generate readings or mage chat for you, because sending your inputs to our AI partner is essential to those features.

This matches the in-app consent screen: your reading and chat inputs are sent to our AI partner Anthropic to write the reply; we never sell your readings or chats; and you can withdraw consent and delete everything in Settings.

14. Children

To reiterate section 3: Arkanika is for adults 18+. It is not directed to children, we do not knowingly collect data from anyone under 18, and we will delete any account we learn belongs to a minor. Because the app is adults-only, California's under-16 opt-in rules for sale/sharing do not apply to our intended users.

15. Deleting your account & data (in-app and web)

You can delete everything at any time: - In-app: Settings → Account → "Delete Account & Data" (two taps) → we soft-delete immediately, then permanently purge your data within 30 days. - On the web: you can request deletion at arkanika.net/delete-account without opening the app (for accounts linked to Apple or Google sign-in — an anonymous guest account has no email we could match to it, so it can only be deleted from inside the app). - Export: you can export your data from Settings before deleting.

When you delete: your readings, chat, display name, economy and account data are purged; if you linked Apple sign-in, we call Apple's token-revocation; any copy held by our AI partner auto-deletes within 30 days (and is never used for training). We may retain de-identified purchase/financial records where the law requires (tax/audit), unlinked from your identity.

When you request deletion through the web page, we keep a minimal record of the request — the email address you enter, any optional note you add, and the request's status — for 180 days, as an audit trail that your request was honoured and to prevent abuse of the form; this record survives the account purge. Deleted account data may also persist in our routine database backups for up to 30 further days before those backups automatically expire; backups are stored privately, encrypted at rest, and are not used to restore deleted accounts.

16. Consistency with our store privacy labels

The disclosures here are designed to match the Google Play Data safety form and the Apple App Privacy ("nutrition") label — the data types we collect, how we use and share them, our security practices, and deletion. If we change what we collect or which third parties we use, we update this policy, the Play Data safety form, and the Apple label together.

17. About the AI, and AI safety

Spread readings, deep readings and mage chat are AI-generated; daily-card readings and standard card meanings are pre-generated in advance (some with the help of our AI partner) and served identically to all users — no personal data is used or sent to the AI for them (see Section 5 of our Terms of Use). All reading content may be inaccurate, inconsistent, or unexpected, and are provided for reflection and entertainment only. They are not professional medical, legal, financial, or psychological advice, and you should not rely on them for important decisions.

  • Every reading and mage message includes an in-app Report control so you can flag offensive or harmful content. When you report content, it is sent to us and a person on our team may review it (see section 4, "Content you report to us").
  • If you are in crisis: If you are thinking about harming yourself or are worried about your safety or someone else's, please contact your local emergency services or a crisis/suicide-prevention helpline right away. You can find helplines for many countries at findahelpline.com. Arkanika is an entertainment app, not a crisis service, and its AI mage cannot keep you safe — a person can.
  • There is no solely-automated decision that produces legal or similarly significant effects about you.

The full AI disclaimer and the same crisis resource also appear in our Terms of Use.

18. Changes to this policy

We may update this policy as the app and the law evolve. We'll revise the effective date above and, for material changes, give notice through the app, a "What's New" note, or email where we have it. Where a change affects processing that relies on your consent, we'll ask for your consent again. Continued use after the effective date means you accept the updated policy, to the extent permitted by law.

19. Contact us & how to complain

If you're in the EEA or UK, you also have the right to complain to your local supervisory authority (or the UK ICO). We'd appreciate the chance to address your concern first.

This policy is interpreted under the laws of Ukraine, without limiting any mandatory rights you have under your local law.