Photo Whisper AI — Privacy Policy
Effective date: 16 Nov 2026
Last updated: 16 Nov 2026
This policy explains how EZA Labs SRL processes personal data when you use Photo Whisper AI (the "App"), visit photowhisper.ai (the "Website") or contact us: what we collect, why, who receives it and your rights.
1. Who we are
EZA Labs SRL, registered in Romania, operates Photo Whisper AI. We are the data controller: we decide what personal data is collected and how it is used.
For privacy questions or requests, email privacy.photowhisperai@ezalabs.io. Full company and contact details appear in Section 10.
2. What data we collect and why
2.1 When you create an account
- Email address — so you can sign in, receive transactional emails, and recover your account.
- Name (optional, only if your sign-in provider shares it) — shown in the App as your display name.
- Apple or Google account identifier — received from your sign-in provider to link your account to that method.
- Authentication and session information — to keep you signed in and protect account access.
We do not routinely ask you for an age band, date of birth or identity document at signup. Our eligibility rules and handling of known underage use are explained in Section 6.
Legal basis: performance of the contract (Art. 6(1)(b) GDPR) for information necessary to create and operate your account; legitimate interests in securing account access (Art. 6(1)(f) GDPR).
2.2 When you use AI features
- Images, video and audio you upload — depending on the feature, stored in our cloud storage and sent to the providers processing your request.
- Text prompts and editing instructions — used to prepare and carry out your request, including prompt enhancement and safety checks where applicable.
- Generated images and videos — stored in our cloud storage so you can access them.
- Processing records — selected features and models, request status, timing, errors and related credit usage, used to deliver the result and resolve problems.
Important: AI processing may take place in the European Union, the United States, and Asia-Pacific, including Malaysia, Indonesia and mainland China, depending on the feature and model. Some video features use a provider's Global inference service: its access point and stated storage region are Frankfurt, Germany, while computation can take place on available nodes worldwide, including mainland China. An EU access point or storage region does not mean that inference stays in the EU. The media and instructions needed for your request are transmitted to the relevant providers. Some providers also perform safety checks, and authorized reviewers may examine flagged content or investigate abuse. Provider retention is separate from storage in your Photo Whisper library (Section 4). We do not sell your content or use it to train our own AI models. Content is encrypted in transit; inference requires providers to process the content itself.
The App asks permission before sharing necessary inputs with third-party AI providers. Decline or withdraw in Settings to stop new requests requiring that sharing. Withdrawal makes those features unavailable; it does not recall transmitted requests or automatically erase retained records. Request erasure separately under Section 5.
Legal basis: performance of the contract (Art. 6(1)(b) GDPR) for processing necessary to provide the AI feature you request, including features supplied with free credits. The App's AI-sharing permission is an additional user control; it is not a blanket GDPR consent to unrelated processing. Safety and abuse-prevention purposes are described separately below.
2.3 When you purchase credits or subscribe
- Purchase and refund information — we receive transaction references, product and store information, price, and purchase, cancellation, or refund status from Apple or Google, both through our subscription-management provider and directly from Apple's App Store Server API and Google Play's records of refunded and cancelled orders. Refund information can include the refund type, the share refunded and the store's reason category.
- Subscription and credit state — whether you're subscribed, on what plan, your credit balance, how much of each purchase you have used, and any adjustment linked to a refunded purchase.
- Refund review information — for each refund: its category (for example, a store refund or a chargeback), the share of the purchase used, the time between purchase and refund, the account's age, other refunds and disputes in the previous 12 months, any account restriction, and whether the refund has been flagged for review.
We do not receive your credit card number, bank account, or any payment credentials. Apple and Google process all payments.
We use this information to provide purchases, prevent duplicate credit grants, reconcile refunds, remove the unused credits of a refunded pack, handle billing support and notify you of credit removals and account restrictions. Where we review suspected deliberate payment abuse, we rely on our legitimate interest in protecting the service and other paying users. Our systems flag a refund, or a purchase made on an account that is on hold, for review automatically, using the information above; a person reviews each flag and decides any account restriction (Section 5).
Store refund reviews. When Google asks us about a chargeback, our subscription-management provider answers for us with our refund preference only. No usage information is sent. We do not answer Apple's refund requests.
Legal basis: performance of the contract (Art. 6(1)(b) GDPR); legal obligation for records we must keep (Art. 6(1)(c) GDPR); and legitimate interests for proportionate payment-abuse review, answering chargeback reviews, and the establishment or defence of legal claims (Art. 6(1)(f) GDPR).
2.4 Terms acceptance and AI-sharing choices
We record acceptance of the Terms and incorporated Acceptable Use Policy, and your AI-sharing permission or withdrawal. Records include your account reference, the document or notice version, the time and, where available, the App version. Acknowledging this Privacy Policy does not mean consenting to all processing described here.
Legal basis: performance of the contract for administering access to requested features (Art. 6(1)(b) GDPR), and legitimate interests in documenting your choices and establishing or defending claims (Art. 6(1)(f) GDPR). Where a particular record is required by law, we rely on that legal obligation (Art. 6(1)(c) GDPR).
2.5 Communications, support and feedback
We currently send service messages, including sign-in links, support responses and important account or policy notices, such as notice that credits from a refunded pack were removed or that an account restriction started or ended. Apple or Google may separately send purchase messages. Essential notices are separate from optional marketing and cannot be turned off while you continue using the affected service.
If we introduce marketing emails, we will use your email address and communication preferences for that purpose only with your valid prior consent. Each marketing email will include an unsubscribe link, and you can also contact us to withdraw consent at any time. Declining or withdrawing marketing consent will not affect your access to the service.
For support, we process your email, message and information needed to resolve your request. In-app feedback and reports include your rating or comment and the generation reference. Investigations may involve related account, request or content information. Content-safety reports also follow Section 2.7. Avoid unnecessary personal information or sensitive attachments.
Website email-preference links use the contact reference, associated email and chosen preference to display or update that choice.
App notifications. If you allow notifications, we let you know when something you started in the App is ready or could not be made while you are not using the App. We do not send promotional notifications. A notification contains a short pre-written message about the result, such as "Your video is ready", and internal references the App uses to open it. It never contains your prompts, images, videos or the AI model used. Depending on your device settings, notifications may appear on your lock screen.
To deliver notifications, the App obtains a push token for your device through Apple (iOS) or Google (Android) and our notification-delivery provider, and creates a random notification identifier for this installation. We store these with your account and sign-in session, together with your device platform. Our notification-delivery provider and Apple's or Google's notification service process the token, related installation identifiers and each notification in order to deliver it. We do not use these identifiers for advertising or tracking.
You can turn notifications off at any time in the App (Settings → Notifications), which removes this device's registration from our systems, or in your device settings.
Legal basis: performance of the contract for account and service support (Art. 6(1)(b) GDPR); legal obligations for required notices and rights requests (Art. 6(1)(c) GDPR); legitimate interests in handling other enquiries, feedback, communication preferences and disputes, and in notifying you about requests you started (Art. 6(1)(f) GDPR); and consent for optional marketing if introduced (Art. 6(1)(a) GDPR). If we introduce promotional notifications, we will ask for your consent first.
2.6 Security and abuse prevention
To keep the App secure, we collect limited technical data:
- IP address and basic device information when you connect to our servers (logged temporarily)
- Approximate city and country inferred from connection metadata when you sign in, included in the account-security email sent to your account address. The App does not request GPS or precise-location access for this purpose
- Activity patterns that help us detect abuse, such as automated scraping or attempts to generate prohibited content
Fraud-prevention identifiers. To prevent fraud and repeated claims of one-time promotional credits, we use account information and pseudonymised (hashed) identifiers derived from your email address, your Apple/Google sign-in ID and a device identifier. On iOS, the device identifier is a random value stored in your device's secure Keychain; on Android, it is derived from your device's Android ID. It normally stays the same if you reinstall the App. We do not use these identifiers for advertising or tracking. The pseudonymised identifiers may remain after account deletion and are scheduled for deletion 360 days after collection, through daily cleanup (Section 4).
Crash and error diagnostics. Release builds send diagnostic reports to our error-monitoring provider to help us investigate crashes and reliability problems. Reports can include an internal account identifier, App and operating-system versions, device information and technical error details. We apply filters to reduce personal data in reports, including filters for email addresses, access tokens and URL query values. Technical messages and request paths can still contain personal information. Operational alerts can also include error details and account or request references. Billing alerts about refunds and purchases can also include your email address, purchase details and refund review information (Section 2.3). We do not enable session replay or advertising tracking through the diagnostic service.
Legal basis: legitimate interests (Art. 6(1)(f) GDPR) in protecting accounts, preventing payment and promotional abuse, and diagnosing service faults. You may object as explained in Section 5. Pseudonymisation does not make these records anonymous.
2.7 Content safety and abuse review
We and our providers use automated safety checks for prohibited requests and content. Checks and assessed data vary by feature and model. Inputs may be checked on our hosting infrastructure before inference; inference providers may also check inputs or outputs.
- Data used: the inputs needed for the relevant check, the result or reason category, and associated request records. Checks can include transcribing uploaded speech and assessing whether an image or prompt appears to show a recognisable public figure or a character from an established franchise, so that the request can be blocked. These checks do not identify private individuals. Review of a report or suspected abuse may require access to the relevant content.
- Effect: a request or result may be blocked automatically. Checks can make mistakes and are not a guarantee that all prohibited content is detected.
- Retention: library content, operational records and separate provider safety logs have different retention rules (Section 4). Evidence may need to be preserved for an investigation or legal obligation.
- Review: contact support.photowhisperai@ezalabs.io if you believe a block or account restriction is mistaken. When reviewing the decision, we may withhold details that would enable circumvention or compromise someone else's safety only to the extent permitted by law.
We do not use the App to assess your eligibility for employment, credit, insurance or similar decisions. A request-level safety block is intended to restrict a particular use of the service. Your rights concerning automated decisions are explained in Section 5.
Legal basis: legitimate interests (Art. 6(1)(f) GDPR) in preventing unlawful or harmful use and protecting users and the service. Where a specific legal duty requires preservation, reporting or disclosure, that processing is based on the applicable legal obligation (Art. 6(1)(c) GDPR).
2.8 Privacy practices and sensitive information
- We don't track you across other apps or websites.
- We don't use advertising identifiers.
- We don't sell your personal data.
- We don't sell your content, and we don't use your uploaded images or prompts to train our own AI models.
- We do not ask you to upload health records, financial account details or government identity documents. Do not include them in content or prompts; if you choose to include information in an upload, it will be processed as part of that upload.
- We do not use your images or voice to establish your identity through biometric identification. Media can nevertheless contain personal information and reveal sensitive characteristics.
- For some features, such as Replace, the App uses Apple's or Google's on-device tools to find where faces and people appear in a photo. The App sends us the position of those areas with your photo, to frame your result. We do not create face templates or use this to recognise anyone.
2.9 When you visit the Website
Hosting providers receive connection and request information, including IP address, browser information and requested pages or resources, to deliver and secure the Website. Linked app stores and external services apply their own privacy notices. Website support and email-preference requests follow Section 2.5.
Legal basis: legitimate interests (Art. 6(1)(f) GDPR) in making the Website available, maintaining its security and resolving technical problems. This basis does not replace consent where applicable law requires it for access to or storage of information on your device.
3. Who we share data with
We share data with the recipient categories below for the stated purposes. Processors acting for us require applicable data-protection terms. Other recipients, including Apple and Google for their own sign-in, payment and notification services, apply their own privacy notices. We do not sell personal data.
| Category | Purpose | Location |
|---|---|---|
| Cloud hosting, databases and storage | Hosting account and processing records, uploaded media and outputs; delivering the Website and operating technical and safety services | App storage in the European Union; Website delivery, processing and administrative access on a global network, including the United States |
| AI inference and content-processing providers | Processing the images, video, audio and instructions needed for your chosen feature; prompt preparation and content-safety checks where applicable | European Union, United States, and Asia-Pacific, including Malaysia, Indonesia and mainland China; some video inference uses available nodes worldwide, with a Frankfurt access point and stated storage region |
| Push-notification delivery provider | Delivering App notifications to your device | United States |
| Email delivery and support service providers | Delivering service messages, including sign-in security details, and handling support correspondence; optional marketing emails if introduced with your consent | European Union and, depending on the communication service, international processing or access |
| Crash, error-monitoring and operational-alert providers | Diagnosing technical failures, investigating incidents and alerting us to refunds and purchases that may need billing review, using technical reports, account, purchase or request references and, for billing alerts, your email address and refund review information | European Union and international processing or access, including the United States |
| Subscription and purchase management | Validating Apple/Google receipts, managing your subscription state and answering Google chargeback reviews with our refund preference | United States |
| Apple and Google | Sign-in with Apple / Sign-in with Google, in-app purchases and refunds, and delivering App notifications | Worldwide |
| Professional advisers and competent authorities | Advice, legal claims, regulatory duties and lawful requests, where necessary | Relevant adviser or authority's jurisdiction |
Personal data may be transferred outside the European Economic Area, including to the United States, Singapore, Malaysia, Indonesia and mainland China, including when a provider forwards requests from an EU access point for Global inference. Depending on the provider and transfer, contractual or adequacy safeguards may include Standard Contractual Clauses approved by the European Commission or the EU-US Data Privacy Framework for transfers covered by a US recipient's certification. For AI inference transfers to Singapore, Malaysia, Indonesia or mainland China, the contractual framework uses Standard Contractual Clauses and applicable onward-transfer obligations; these countries do not have an EU adequacy decision, and the EU-US Data Privacy Framework does not cover this processing.
For transfer details or a copy of the safeguards, email privacy.photowhisperai@ezalabs.io.
3.1 When we disclose data to authorities
We may disclose relevant data in response to a lawful obligation or request, or where another applicable legal basis permits necessary and proportionate action to protect someone's rights or safety. We assess the request and applicable disclosure and transfer requirements; an overseas request does not by itself remove those requirements.
4. How long we keep your data
| Data | Retention |
|---|---|
| Account and authentication data | While needed to operate your account. Account deletion closes access and starts removal or de-identification of profile and related records; limited records needed for billing, security or claims follow the criteria below |
| Images and videos in your Photo Whisper library — while you have an active subscription | Retained while your subscription is active, unless you delete them or your account, or removal is required for another lawful reason |
| Images and videos in your Photo Whisper library — without an active subscription | 10 days from creation for new media or from the end of subscription protection for previously protected media, then scheduled for deletion |
| Related prompts, reference media and processing records | While needed to carry out the request and provide the associated feature or library item; deletion removes or de-identifies associated personal content. Limited billing, security or claims records may need separate retention |
| Prompts and responses processed by the Global video-inference provider described in Section 2.2 | Normally retained by that provider for up to 30 days; longer where required by law or reasonably necessary for service security. Asynchronous task files are stored for processing and delivery, and separately stored files follow the provider's feature-specific retention rules |
| Inputs and outputs that trigger AI inference providers' content-safety filters | Depending on the provider, flagged-content logs may be retained for 180 days in Malaysia, or flagged content for up to 2 years by the Global video-inference provider for law-enforcement cases and rights-protection procedures, separately from your Photo Whisper library |
| Accounting documents we are legally required to retain | Generally 5 years calculated from 1 July of the year following the relevant financial year, under Romanian accounting rules; a specific legal obligation or pending proceeding may require longer retention |
| Other purchase, refund and credit records | While needed to administer your paid entitlements, reconcile transactions and resolve disputes; relevant evidence may be retained through the applicable claim or limitation period |
| Refund review information (Section 2.3) | 12 months from the refund, or until account deletion if sooner. The refund record itself follows the purchase-records rule above |
| Account restriction records (reason, evidence, notices and outcome) | While the account exists and, after that, only while needed for a related claim, normally no longer than 3 years after the restriction ends |
| Content-safety check records (Section 2.7): the checked prompt, the result and its category and, when a check does not pass, the check's explanation and any transcript of uploaded speech. They contain no images | Kept with the related request records. The prompt, explanation and transcript are erased when you delete the related image, video or creation, or your account; records of a blocked request with nothing to delete may therefore stay until account deletion, to handle a review of the block and detect repeated misuse. The result and category may be kept without that content as a security record |
| App and Website security records, including sign-in connection and approximate-location details, moderation event logs, crash and error records | For as long as necessary to diagnose faults, protect account access, investigate incidents and detect repeated misuse, considering the record's sensitivity and the duration of the issue; evidence relevant to a legal claim or preservation duty may be retained for that matter |
| Pseudonymised fraud-prevention identifiers (email, Apple/Google sign-in ID and device identifier — see Section 2.6) | May remain after account deletion; scheduled for deletion 360 days after collection through daily cleanup, to prevent repeat claims of one-time promotional offers |
| App notification registration (push token, notification identifier, device platform) | Until you sign out, turn notifications off in the App, delete your account or uninstall the App. If the App stops connecting while you remain signed in, within 8 days of its last connection. Notification services may briefly hold a notification until it can be delivered |
| Terms-acceptance, AI-sharing permission, support, feedback and communication-preference records | While relevant to your account and, where necessary, through the applicable period for demonstrating your choices, resolving a request or establishing or defending a claim |
Library retention does not govern provider retention. Providers may retain request or output data for delivery, troubleshooting or abuse prevention under the feature's terms. The 180-day and 2-year periods concern particular flagged content, not all requests. A delivery link expiring does not establish that every provider copy or log has been deleted. Library or account deletion does not automatically erase provider logs. Request erasure or request-specific retention information under Section 5.
Deletion can be asynchronous. Limited records may remain for the purposes above, and backups follow their retention lifecycle. Account deletion does not make every recipient's copy disappear immediately. Retained data remains subject to data-protection duties and your rights.
5. Your rights
Under GDPR, you have the right to:
- Access — ask what personal data we hold, receive a copy, and obtain information about processing and recipients, including the actual recipients' identities where required by law.
- Rectification — correct data that's wrong or incomplete.
- Erasure ("right to be forgotten") — request deletion of your data. You can request account deletion directly in Settings or by emailing privacy.photowhisperai@ezalabs.io, including when you cannot log in or no longer have the App installed. Exceptions and retained records are described in Section 4.
- Restriction — ask us to pause certain processing.
- Portability — receive your data in a structured, machine-readable format.
- Object — object on grounds relating to your situation to processing based on legitimate interests. We must then stop unless the applicable legal grounds for continuing are met. You also have an unconditional right to object to direct marketing. You can stop App notifications at any time in Settings → Notifications or in your device settings.
- Withdraw consent — where processing relies on GDPR consent, withdraw it at any time without affecting earlier lawful processing. AI-sharing permission can separately be withdrawn in Settings as explained in Section 2.2.
- Automated decisions — where Article 22 GDPR applies, you have protections against decisions based solely on automated processing that produce legal or similarly significant effects, including applicable rights to human intervention and to contest a decision. When a store confirms a refund of a credit pack, the pack's unused credits are removed automatically, as our Terms provide. Payment-abuse flags are automated, but a person decides every account restriction. To contest a credit removal, a block or an account restriction and have a person review it, contact support.photowhisperai@ezalabs.io.
How to exercise these rights: email privacy.photowhisperai@ezalabs.io. We respond without undue delay and normally within one month. If complexity or the number of requests permits an extension, we will explain the reason within that first month; the extension can be up to two further months. We may need proportionate information to verify your identity. Rights are subject to their legal conditions; if we cannot act, we will explain the reason and available complaint options. Requests are normally free of charge.
Right to complain: if you believe we've mishandled your data, you can lodge a complaint with the Romanian data protection authority (ANSPDCP — Autoritatea Națională de Supraveghere a Prelucrării Datelor cu Caracter Personal, https://www.dataprotection.ro), or with a competent supervisory authority, including in the EEA country where you live or work or where an alleged infringement occurred.
United Kingdom: where UK data-protection law applies, the rights described above also apply subject to that law's conditions. You can complain to the Information Commissioner's Office at https://ico.org.uk/make-a-complaint/.
Additional US state rights: subject to the applicable law's scope, conditions and exceptions, you may also have rights to access, correct, delete or obtain a portable copy of personal data, and appeal a refusal. Send requests and appeals to privacy.photowhisperai@ezalabs.io; we explain any applicable appeal route when refusing a request. Where required by law, authorized agents may act for you with proportionate identity and authority verification. We will not discriminate against you for exercising applicable privacy rights.
6. Who can use the App
The App is for people aged 18 and over who have reached the age of majority where they live and have contractual capacity, as explained in the Terms' eligibility requirements. Store ratings are separate and do not establish a user's age. This policy is a processing notice, not consent to every processing activity.
We do not routinely collect age bands, dates of birth or identity documents at signup, and we do not claim to verify each user's age. If we receive reliable information that an account holder does not meet our eligibility requirements, we investigate proportionately and restrict confirmed ineligible use or close the account as appropriate. We do not automatically infer age from uploaded faces. Our retention rules and applicable privacy rights continue to apply.
Parents, guardians and users can report underage accounts or make related rights requests to privacy.photowhisperai@ezalabs.io. Provide only the identifying information needed to locate and review the account.
7. Security
We use technical and organisational measures designed to protect your data, including:
- Encryption in transit (TLS) and at rest
- Access controls and logging on our production systems
- Use of reputable infrastructure providers with recognised security programmes (such as SOC 2)
- Contractual security obligations with our service providers where a data-processing agreement is in place
No system is perfectly secure, but we act quickly if we detect a problem. If a personal-data breach creates a high risk to your rights and freedoms, we'll notify you without undue delay as required by Article 34 GDPR. We'll also notify the Romanian data protection authority within 72 hours where required by Article 33 GDPR.
8. AI-specific transparency
- Outputs use AI. The App's AI features create or modify images and videos using artificial intelligence. Outputs can resemble real recordings. Images and videos you save or share from the App carry standard machine-readable AI-origin metadata (IPTC Digital Source Type) and normally a visible AI mark. The metadata identifies the file as made or edited with AI in Photo Whisper; it contains no account identifier or prompt. Keep these markings when sharing an output; they do not establish that a depicted event actually happened.
- AI outputs can be wrong. They may be inaccurate, unrealistic, or inappropriate. Don't rely on them for anything legally, medically, or financially consequential.
- Only upload yourself when a person is identifiable. You must have the necessary rights in your upload, and any identifiable real person depicted or heard must be you. Another person's permission does not override this rule. Content without an identifiable person is allowed subject to our rules.
- We do not verify identity from your media. Safety checks are not an identity-verification service and do not confirm that an upload depicts its sender.
- Safety checks vary by feature. Automated checks and review of reported or suspected misuse are described in Section 2.7. Authorized provider review for safety purposes may also occur.
- Our Acceptable Use Policy lists prohibited content and uses. Reports, rights-holder complaints or legal orders may alert us to violations. Removal and access restrictions follow the AUP and Section 16 of the Terms of Service; we cooperate with authorities where required by law.
9. Changes to this policy
We update this policy for changes in processing or legal obligations. Material changes affecting data use or your rights receive clear notice by email or in the App; Website-only changes may be notified there.
For new purposes or materially different processing, we provide information before processing begins, allowing time proportionate to the impact to understand the change and exercise your rights. We obtain any legally required consent before processing; continued use does not replace consent. Corrections and clarifications that do not materially change processing or rights may take effect on publication.
Existing protections under earlier policy versions, our Terms or law remain applicable, including any 30-day notice commitment and its stated exception. This revision does not retrospectively shorten an applicable notice period.
The current published English version is available at https://photowhisper.ai/privacy-policy/, with its effective date and last updated date.
10. Contact
Questions, requests, or complaints:
EZA Labs SRL
Str. Brașov 18, Bl. F1, Sc. 4, Et. P, Ap. 56, 061447 București, Romania
Trade registry: J2021013958409 (EUID: ROONRC.J2021013958409)
Email: privacy.photowhisperai@ezalabs.io