Adam & Eva

Privacy Policy

2026-10-16-v6

1. Who controls your data

The controller is Pavlo Olehovych Konoshchenko, operating Adam & Eva in Ukraine as a sole proprietor (FOP). Privacy and data-protection enquiries may be sent to support@adam-eva.app. No data protection officer (DPO) is designated.

This Policy explains the processing implemented by the Adam & Eva website and Android and iOS applications. It is information about processing; it is not a request to waive privacy rights or give blanket consent.

2. Service and age

Adam & Eva is an 18+ faith-oriented dating and communication service. We do not knowingly permit anyone under 18 to create an account. If you believe a minor uses the service, contact us so we can investigate and take proportionate action.

3. Data we process

4. Why we process data and our legal bases

Where consent is the basis, refusing it does not retroactively affect other lawful processing. Contract and required account/security data may be necessary to provide the basic service; we explain the consequence when a field is optional.

5. What other users can see

Depending on profile configuration and your settings, other users may see your name, age, main and additional photos, country/city, online or last-active status, biography, faith/confession and other profile answers. A health-information choice is exposed only as a privacy-safe indicator that you wish to discuss something before meeting, not as diagnosis details. Likes, matches, messages and private support/moderation material are not public. Invisible/privacy settings and blocking rules restrict visibility but cannot remove data another person already saved outside the service.

6. Discovery, ranking and profiling

The service filters and ranks profiles using information such as age, sex, country, profile completeness, preferences, activity, Premium/promotions and safety/privacy eligibility. It also derives aggregate map counts. These processes recommend who appears; they do not make a solely automated decision producing legal or similarly significant effects within the inspected implementation. Moderation or account restrictions involve configured rules and/or authorized review. Contact us if you want meaningful information about a result affecting your account or wish to contest it.

7. Messages, photos and moderation

Messages and attachments are stored to deliver conversations, synchronise devices and apply blocks/read state. Authorized personnel may access relevant content when a user reports it, support is requested, safety or fraud is investigated, or law requires it. We do not promise that every conversation is proactively read.

Profile photos are stored in Microsoft Azure Blob Storage and processed locally for resizing. Photo bytes may be sent to Microsoft Azure OpenAI for safety-policy classification and to Azure Face to detect a face rectangle/presence. Azure Face is configured with `returnFaceId=false` and performs no provider-side identity matching. Separately, optional Photos verification derives transient numerical face descriptors (embeddings) within the service and compares the consented selfies with approved profile photos for same-person similarity; this does not establish legal identity. Moderators may review photos and decisions. Future biometric identification or a materially different AI purpose requires a new assessment and, where applicable, new consent.

8. Payments, gifts and subscriptions

Stripe processes Web checkout/subscriptions; Google Play processes Android purchases and subscriptions. They receive the data required to perform and verify the transaction. Adam & Eva stores processor identifiers and entitlement/status records, not complete card details. The payment provider may act under its own legal duties for payment, fraud and accounting. Purchases are also governed by the checkout/provider terms and mandatory consumer law.

9. Notifications, support and security

Firebase Cloud Messaging receives push tokens and delivery payload metadata to deliver notifications. Azure SignalR supports real-time updates. Anonymous support requires an email so support can understand and respond to the request; creating a fake account is not required and we do not use that email to reveal whether an account exists. The basis is performing your requested support step and our legitimate interest in complaint handling and security, not marketing consent.

We use authentication/session data, rate limits, IP/network signals, moderation records, logs and alerts to prevent abuse and diagnose incidents. Acceptance evidence records the authenticated account, exact legal version/hash, server UTC, platform/client version, session and correlation ID. It deliberately avoids aggressive device fingerprinting and full request bodies.

10. Recipients and service providers

Data is disclosed only as needed to other users, authorized support/moderation/operations personnel, payment providers and these evidenced providers:

We may disclose data to competent authorities or professional advisers only where lawfully required or necessary for rights, safety and legal claims. We do not describe an unverified contractor or analytics provider as a recipient.

11. International transfers

Core Azure application/SQL/storage resources inspected are in West Europe, while configured Azure Face and Azure OpenAI resources are in East US; global providers can also process operational data outside the EEA. A technical resource location alone does not prove the contractual transfer safeguard. For a restricted transfer, the controller must ensure an applicable adequacy decision, approved contractual clauses or another lawful GDPR Chapter V mechanism and supplementary measures where needed. You may request information about the mechanism relevant to your data. The controller must verify the current provider agreements before relying on a named safeguard.

12. Retention

We do not retain an entire deleted profile “just in case.” A documented legal hold may suspend deletion only for the narrow data and period necessary.

13. Account deletion

Use the in-app account-deletion control or contact support. The workflow removes the active profile, photos/blobs, chats, likes, matches, device/push/auth records and other account-linked product data. Backups may persist temporarily until overwritten under the provider schedule. Minimal security, accounting, moderation or legal-evidence records may be retained only when a real legal obligation, safety need or establishment/exercise/defence of claims justifies it, with restricted access. Legal acceptance is never backfilled from account creation and deletion does not rewrite historical evidence.

14. Your rights

Subject to the GDPR and applicable limits, you may request access, correction, erasure, restriction, portability, and information about processing; object to legitimate-interest processing; withdraw consent at any time as easily as it was given; and contest a relevant automated or moderation outcome. Withdrawal does not make earlier lawful processing unlawful. We may need proportionate verification and must explain any lawful refusal or extension.

Contact support@adam-eva.app. You may complain to the supervisory authority where you live, work or believe an infringement occurred. For the Ukrainian controller, you may also complain to the Ukrainian Parliament Commissioner for Human Rights. You may also seek a judicial remedy.

15. Security and incidents

We use access controls, authenticated sessions, encryption in transit, restricted administrative roles, rate limiting, content moderation, audit/correlation records, token protection and provider security controls. No online service is risk-free. If a personal-data breach creates a legal notification duty, we will notify the relevant authority and affected users as required without undue delay.

16. Changes

Each published Policy keeps a fixed version, exact content hash, publication date and effective date. Old versions are retained. A notice update is not treated as blanket consent. If a new purpose needs consent—especially Article 9 consent—we will request it separately before that processing. Material Terms changes are handled through the Terms acceptance flow.

Illegal-content notices

If you use the public form to report potentially illegal content, we collect the email and name you provide, the exact electronic location of the content, your explanation of why it is illegal, additional information, and your good-faith accuracy declaration. We use these data to assess the notice, communicate about it, protect affected users, meet applicable platform-law duties, and establish, exercise or defend legal claims. The legal basis is the requested complaint-handling step, compliance with applicable legal obligations, and our legitimate interests in platform safety and claims. The notice is retained only for the case and a proportionate complaint, safety and legal-claims period; access is restricted. A notice does not itself cause an automatic restriction.

Consent evidence and marketing preferences

To prove a legal acceptance or consent choice, we keep an append-only event linked to the authenticated account with server UTC, exact document and statement versions and hashes, locale, platform, app/build version, authentication method, session and request ID, a limited user-agent string and the network address protected at rest. We do not store passwords, tokens or persistent hardware identifiers in this record. Marketing email and promotional app push are sent only while a separate voluntary marketing consent is active. A missing checkbox during a later login does not withdraw an earlier consent; withdrawal is available in Settings. Security and transactional messages remain separate.

Safety, fraud prevention and metadata-only Shadow Mode

For the legitimate interests of protecting users, preventing and detecting fraud, scams and abuse, protecting service integrity and improving human moderation, Adam & Eva uses a limited internal metadata-only safety system. It evaluates account creation/login/activity timing, bounded activity and first-contact rates, the existence and frequency of reports and blocks relative to meaningful contacts, and consistency between the profile country and a reliable country already derived from the connection. Country inconsistency is only a weak supporting signal; it is not nationality scoring or proof of wrongdoing.

This internal profiling runs in Shadow Mode. It creates internal safety events, an explainable composite risk indicator, snapshots and a human-moderation queue, but it does not show warnings, rank or hide profiles, restrict, suspend or ban accounts, or otherwise change access to the service. It does not inspect private-message or report text, photos, audio or video; does not use exact GPS; and does not use religion, ethnicity, health, sexual orientation, political opinions or other special-category data as risk factors. The indicator is not a probability of fraud or a criminal finding.

Derived Shadow events, metrics, profiles and scores are retained for no more than 90 days. Admin access-audit records are retained for no more than 365 days. Closed report/safety cases are retained for no more than 12 months after closure unless a specific documented legal or security hold applies. Subject to applicable law, you may request access, correction, restriction, erasure or human review and object to processing based on legitimate interests by contacting support@adam-eva.app.

Photo verification and optional review-evidence retention

Photo verification is an optional trust feature. With your photo-verification consent, Adam & Eva processes exactly three guided camera selfies and the approved profile photos selected for that attempt, including transient numerical face descriptors, to assess same-person similarity. It is not KYC, does not verify a legal name, age, government identity or documents, and is not bank-grade liveness or proof of legal identity.

`photos-verification-manual-review-v2` is a separate optional consent for retaining review evidence. Its control is unchecked by default. If you select it before submitting, Adam & Eva temporarily stores one private evidence bundle containing exactly the three submitted selfies and protected evidence copies of the exact approved profile photos recorded for the attempt. Both types of copies use the same retention deadline and deletion controls; configured retention can never exceed 72 hours. They are not public or used to train models.

Only authorized moderator or administrator roles may access the bundle, and access and decisions are audited. Evidence from a NO_FACE result may be inspected only to diagnose the capture and guide a retry; it cannot support manual approval. A moderator may correct a visual false rejection only for LIVE_FRAMES_MISMATCH—when every retained selfie passed face-presence, capture-quality and required-pose checks—or PROFILE_PHOTO_MISMATCH, and only while the complete exact bundle remains available.

Manual approval does not rewrite the automated FAILED result and is not a KYC or legal-identity determination. A separate audited approval may grant Photos verified status for this limited trust feature, and you are notified.

If you do not select the separate consent, no review-evidence bundle is retained and manual review is unavailable; the submitted selfies are otherwise processed transiently for the automated attempt. You can delete a retained bundle early or withdraw photo-verification consent using the available account controls, and account deletion uses the same active-deletion path. Active selfie and protected profile-evidence copies are removed together; provider backup or soft-delete copies may remain only until their configured expiry with restricted recovery access.

AI Assistant for safer Christian dating

The controller is Pavlo Olehovych Konoshchenko, operating Adam & Eva in Ukraine as a sole proprietor (FOP). support@adam-eva.app.

Get private guidance on your conversations: what to notice, how to express boundaries and continue getting to know someone with respect for Christian values.

Either participant may request an analysis. Azure AI receives chat text, captions, link text, and both participants’ age, bio and photo-verification status. Attachments are not analyzed and links are not opened. Only the requester sees the result and private follow-ups. AI can be wrong and does not guarantee safety. Analyses are kept for at most 90 days.

Permission is voluntary, belongs only to this conversation and applies across your account’s devices. It covers the available previous messages of this conversation, including messages sent from older apps after permission is granted. Both participants must explicitly accept the current rules using an initially empty checkbox; previous global permission, opening an invitation or reading the rules is not acceptance. A request to enable the assistant is a separate preference and does not start analysis. Declining does not indicate risk. Ordinary messaging remains available when someone declines or withdraws, when an older client is used, or when AI is unavailable. The separately controlled mandatory messaging-block mode is not enabled under this publication.

Chats can contain religious beliefs, health, sexuality and information about other people. The permission covers your own permitted text and profile information, including sensitive information you choose to share. Both participants’ permission does not authorize disclosure of third parties’ sensitive information. Do not include unnecessary third-party details or secrets. Minimization reduces obvious identifiers but cannot guarantee their complete removal.

The service provider is Microsoft Azure AI under our accepted Microsoft Customer Agreement and its incorporated Data Protection Addendum (DPA), not a direct API of an external model provider. The resource is in East US and uses GlobalStandard: inference may occur in any geography where the model is available, not only in the EU or East US. Data at rest, including the abuse-monitoring store, stays in the resource's designated Azure geography. The DPA includes contractual transfer safeguards, including Standard Contractual Clauses for covered transfers; you may request information and a copy through support@adam-eva.app. These safeguards do not make processing EU-only.

Azure uses automated abuse review by default; prompts and outputs are not stored by that automated review. Content flagged for possible abuse may instead be retained for review by authorized Microsoft employees. We have no confirmed exemption from this storage or human review. Microsoft's current terms describe temporary storage for investigating and preventing abuse. We have not verified a published numerical maximum or an account-specific deletion timetable for these copies. We cannot promise zero provider retention, a 30-day limit, or immediate provider deletion when you withdraw permission.

Permission does not run in the background, send replies, report anyone or impose sanctions. Either participant may request a private analysis only while both current permissions and other safeguards allow it. New messages do not automatically update an analysis. Moderators’ aggregate statistics do not grant access to private analyses.

The 90-day maximum runs from the original analysis and includes follow-ups, translations and summaries; further questions do not extend it. You can delete an analysis earlier. Deleting the shared conversation or withdrawing either participant’s permission removes affected derivatives for both owners, without deleting chat messages merely because permission was withdrawn. Account deletion and source-message edits/deletion also invalidate relevant derivatives.

Either participant may withdraw permission in this chat’s assistant settings. This prevents new processing and deletes both participants’ related analyses, private AI conversations, translations and summaries, without deleting ordinary messages. Pending work is invalidated; a late provider response cannot restore it. A fresh generation requires new explicit permission from both participants before AI resumes. Each participant controls only their own request to enable the assistant. Azure requests already sent and their costs cannot be recalled.

Minimal financial records contain no conversation content and are stored separately. Their link to your account is removed after 90 days or when your account is deleted.

The SQL database backup configuration checked on 2026-10-15 retains short-term recovery copies for 7 days and has no scheduled long-term retention. These restricted copies are not content available in the app. Deletions and permission withdrawals must be reconciled before restored data is exposed. The 7-day backup period and the application's 90-day derivative limit do not describe Microsoft's abuse-monitoring retention. Deleting data in the app does not establish that Microsoft has already erased a flagged copy. Contact support@adam-eva.app to exercise deletion rights; the DPA provides customer deletion rights and Microsoft's assistance with data-subject requests, not a verified flagged-copy deletion timetable.

This purpose is separate from platform moderation, general terms, marketing and photo-verification permission. Minimal evidence records the authenticated account and conversation, server time, permission and disclosure versions and hash, processing scope including previous history, state generation, request ID, platform and entry point. No old global permission is converted into chat permission. For this optional purpose, the selected legal basis is each participant's separate, explicit consent, including for their sensitive information and the described transfer to Microsoft. Publication or acknowledgement of this notice is not consent. Rights and privacy questions: support@adam-eva.app.

AI-permission evidence is separate from the general legal-acceptance ledger and contains no conversation or AI content. Minimal permission/withdrawal evidence is retained for the lifetime of this conversation and both accounts, and is removed when the shared conversation or either account is deleted. Withdrawal alone retains this evidence and the invitation seen/declined choice; it does not retain private AI results. The general policy’s post-account legal-evidence retention does not apply to these AI-permission records. Ordinary messaging remains subject to existing blocks, moderation and communication rules.

Microsoft Azure: Data, privacy, and security · Microsoft Azure: Abuse monitoring · Microsoft DPA