Privacy Policy

This document, referred to as the "Privacy Policy" outlines the terms governing collection and utilization of Personal Information from Users by Ergashev Zarifjon ("Developer", "we", "us", or "our") during download, installation, and use of mobile applications published by Ergashev Zarifjon.

Consent of the User

By downloading, installing, and using the Mobile application, you voluntarily provide written consent for the processing of your Personal Information as specified in this Policy.

1. General Terms

1.1. This Policy constitutes an integral part of the End User License Agreement (EULA) for the Mobile application. By downloading and utilizing it, you fully accept the terms outlined herein.

1.2. Processing methods include collection, recording, systematization, storage, updating, extraction, use, transfer, depersonalization, blocking, deletion, and destruction, with or without automated facilities.

1.3. EULA and other agreement terms apply unless explicitly stated otherwise in this Policy.

2. Personal Information

2.1. Users do not create accounts or provide personal information such as name or email, except optionally during product surveys. Some of our applications collect a device identifier (Device ID) to enable core app functionality and analytics, as described below.

2.2. Personal Information includes the User's email or social media account provided during product surveys for users who made purchases.

2.3. The User acknowledges third-party software use, which may collect and transmit data as described in Section 3.

3. Third-Party Software and Data Collection

Our applications may integrate the third-party services listed below. Not every service appears in every application. Where an application has its own section in this Policy, that section states which of these services it actually uses, and it governs for that application.

• Supabase — our backend infrastructure provider (database, file storage, and serverless functions). Processes a device identifier to enable core app functionality. See Section 4 for how photos are handled.
• RevenueCat — manages in-app subscriptions and purchases on top of Apple's StoreKit. Receives an anonymous app user identifier and purchase history; all payments are processed by Apple.
• Amplitude — product analytics. Collects an application-generated device identifier and product interaction data (such as feature usage and app events) to help us understand how our applications are used and improve them. It is not used for advertising.
• Google AdMob — used in some of our applications (for example, Jilva) to display third-party rewarded ads. Collects advertising data that is not linked to the user's identity and is not used for cross-app tracking.
• Firebase, Google Analytics, Apple Analytics — may collect device information and usage data for analytics purposes, subject to their respective privacy policies.

None of the data collected by these third-party services is used for tracking purposes (as defined by Apple), nor is it shared with data brokers.

4. Face Data and Photo Processing (All Photo Applications)

This section applies to all photo applications published by Ergashev Zarifjon — including Rassom: AI Photo Maker, Jilva: AI Photo Generator, and Memora – Restore Old Photos AI (each, the "App") — and describes how these apps handle face data and user photos.

4.1. What Face Data We Collect

When you use the AI photo generation or photo restoration features of our apps, you may provide a photo containing a face — for example, a selfie in Rassom or Jilva, or an old family photograph in Memora. The App uses Apple's Vision framework to perform on-device face detection to verify that a face is present in your photo. This detection returns only a presence indicator (whether a face is detected or not). No facial geometry, biometric identifiers, face embeddings, or facial recognition data is extracted, created, or stored by this check. Where an App offers optional on-device retouch effects, temporary facial landmark coordinates are computed locally as described in Section 4.5. The App does not perform facial recognition, create face templates, or generate any biometric information as defined under applicable biometric privacy laws.

4.2. How We Use Your Photo

Your photo is used solely to generate the result you request — an AI-transformed image based on the style template or custom prompt you select (Rassom, Jilva), or a restored, colorized, repaired, or enhanced version of your photo (Memora). The photo is compressed to JPEG format and transmitted securely via HTTPS to our server (hosted on Supabase), where it is immediately forwarded to our AI processing providers. Your original photo is never written to our database or file storage systems. It exists only in server memory during the brief processing period (typically a few seconds) and is discarded immediately after being forwarded to the AI provider.

4.3. Third-Party AI Processing Providers

To generate AI images and photo restorations, your photo is transmitted to the following third-party AI providers:

• OpenAI (OpenAI, L.L.C.) — via the Images API for template-based generations and photo restorations. OpenAI's data usage policy applies: https://openai.com/policies/privacy-policy
• Google (Google LLC) — via the Gemini API for certain generations and restorations. Google's privacy policy applies: https://policies.google.com/privacy

These providers process your photo solely to generate the requested AI image. We do not authorize these providers to retain, store, or use your photo for model training or any other purpose beyond fulfilling the generation request.

4.4. Storage and Retention of Face Data

• Your original photo (input): Never stored on our servers. It exists only in device memory during processing and in transit to the AI provider. It is discarded immediately after the generation request is sent. There is no database column, file storage entry, or backup of your original photo on our servers.
• Generated image (output): Temporarily stored on our servers during download (typically seconds to minutes), then permanently deleted from our servers after you download it. The generated image is saved only on your local device under your control.
• On-device face detection data: Processed entirely on your device using Apple's Vision framework. No face detection data is transmitted to our servers or any third party.

4.5. No Biometric Data; On-Device Retouch Effects

Our applications do not collect, store, or process biometric identifiers or biometric information as defined under applicable biometric privacy laws (including but not limited to BIPA and GDPR). Face detection is used solely to confirm a face is present in the photo and does not create a face template, face print, embedding, or any persistent facial geometry data. Where an App offers on-device face retouching features (for example, skin smoothing, teeth whitening, and eye enhancement in the built-in photo editor of Rassom, Jilva, and Memora), these features use temporary facial landmark coordinates computed locally by Apple's Vision framework solely to locate the image regions to adjust. These coordinates exist in memory only while the effect is being applied, are discarded immediately after the editing operation completes, are never stored, are never transmitted off the device, and are never used to identify you.

4.6. Your Rights Regarding Face Data

Since we do not store your original photos on our servers, there is no server-side face data to access, correct, or delete. You may delete any generated images from the app at any time. You may also choose not to provide a photo, in which case the AI generation feature will not be available for use.

5. Exam Prep

This section applies to Exam Prep and describes how that application handles data. Exam Prep is a study application: it imports or generates practice tests and flashcards from material the user provides. It contains no advertising, and the AdMob, Supabase and face-data provisions above do not apply to it. Where this section and the general sections conflict, this section governs for Exam Prep.

5.1. No Account, and What Identifies a Device

Exam Prep has no sign-up, no login and no user account. It never asks for an email address, a phone number or a real name. On first launch the App generates a random identifier for the device (a UUID). That identifier is not derived from any hardware identifier, is not the advertising identifier, and cannot be traced back to a person. It is used to deliver AI results to the device that asked for them, to send push notifications, and to group analytics events from the same installation. Deleting the App discards it; a reinstall produces a new one.

An optional first name may be entered during onboarding to personalize greetings inside the App. It is stored only on the device and is never transmitted to our servers.

5.2. Study Data Stays on the Device

Tests, questions, answers, flashcards, decks, scores, streaks, goals and study history are stored locally on the device, in the App's own database. They are not uploaded, not backed up to our servers, and not readable by us.

5.3. AI Generation

When the user asks the App to generate a test or flashcards, the material they choose — a PDF, a Word document, a video or audio file, or a YouTube link — is uploaded over an encrypted connection to our processing server and forwarded to Google Gemini, which produces the questions. Uploaded files are deleted from the server as soon as the job finishes; they are never added to a database or to permanent storage. Generated results are held for a maximum of 48 hours so the device can collect them, and are deleted automatically after that, or sooner once the device confirms receipt.

Material sent for AI generation is processed by Google under its own terms. Users should not upload material they are not permitted to share, and should avoid uploading documents containing personal or confidential information about other people.

A count of successful generations per billing month is kept against the anonymous subscription identifier in order to enforce fair-use limits. It records how many generations were made, never what they contained.

5.4. Subscriptions

AI generation is a paid feature. Subscriptions are handled by RevenueCat on top of Apple's and Google's own billing systems. RevenueCat receives an anonymous subscriber identifier and the subscription receipt. Payment details — card numbers and billing addresses — are handled entirely by Apple or Google and are never seen by the App or by us.

5.5. Push Notifications

If the user allows notifications, Firebase Cloud Messaging issues a delivery token that is stored on our server alongside the device identifier, so the App can be told when a generation job has finished. Notifications may also include study reminders scheduled on the device itself. Permission can be withdrawn at any time in the operating system's settings.

5.6. Analytics

Exam Prep uses Amplitude to understand how the App is used in aggregate — for example how many people finish onboarding, which features are opened, and where people stop. What is sent is limited to the name of the action, counts, durations, percentages and fixed option values such as the chosen difficulty or interface language.

No content is sent to Amplitude: not test titles, questions, answers, file names, links, goal titles or anything else the user typed. Analytics events are attached to the App's own anonymous device identifier described in Section 5.1. The advertising identifier, the vendor identifier and precise location are explicitly disabled in the App's configuration. An approximate country is derived from the network address so that we can tell which regions the App is used in; no finer location — city, region or coordinates — is collected. This data is not used for advertising, is not combined with data from other apps or companies, is not shared with data brokers, and is not tracking as Apple defines it.

5.7. The Library

The Library lets users download tests published by us and, where the feature is available, submit their own test for others to use. Submitting is entirely voluntary. A submitted test is transmitted to our server and, if approved, published to other users, together with the display name the submitter chose to attach to it. Nothing is published unless the user actively submits it. A published test can be taken down on request using the contact address in Section 15.

5.8. Children

Exam Prep is rated for ages 13 and over. It is not directed at children under 13 and does not knowingly collect data from them.

5.9. Requesting Deletion

Because Exam Prep holds no account, the data held on our servers for a device is limited to the device identifier, its notification token, any generation results not yet expired, the fair-use counter and any Library test the user chose to submit. Deleting the App removes everything stored on the device, including the identifier itself.

Generation results expire on their own within 48 hours. The device record and its notification token remain until deletion is requested. To request deletion, or the removal of a submitted Library test, contact us at the address in Section 15 and describe the submission or the approximate dates of use so that the record can be located; because the identifier is held only on the device and is not linked to any personal information, we cannot otherwise identify which record belongs to a given person.

6. PlayScan — Sharing Scores by Link

In PlayScan you can share a score you scanned by link. Shared scores are not listed anywhere in the app or on the web and are not offered to search engines. This section describes what happens when you share a score and when someone reports one; full details are at playscan.app/legal/privacy.

• What we receive: only the recognized notes of your scan (the MusicXML text of the result), the title and composer you enter, the ID of the scan it came from, your device identifier, the platform and the app version. Your photos, annotations and settings are never sent.
• What is shown: the notes, their playback, the title, the composer and any lyrics. Credits, rights notices and other identification text recognized from the page (except the composer) are removed from the shared copy. Your device identifier is never shown to other people.
• Who can see it: only people who have the link, in the app or on the playscan.app/s/… web pages. Messaging apps may keep their own preview of a link (its title and composer), and we cannot delete it.
• Checking that a score is your own scan: our server keeps a fingerprint (a hash, not the notes themselves) of every result recognized for your device, together with your device identifier, for 365 days.
• How long it is kept: until you stop sharing it (Settings → My shared links) or we remove it. The link then stops working at once and its files are erased within minutes; 30 days later we delete your device identifier, title and composer, and only technical data remain. Database backups may keep a copy for up to 30 more days.
• Reports: anyone who opens a shared score can report it. A report includes the reason, an optional comment and the reporter's device identifier (in the app) or, on the website, a code derived from the IP address and, for rights holders, a name and email address. Reports are deleted 12 months after we resolve them. We may remove a shared score or block a device from sharing.
• If you reinstall the app: you lose control of the scores you shared from it. Send us the link (Section 15) and we will remove it. Copies that other people added to their library stay on their devices but cannot be exported.

7. PlayScan — Suggestions You Send

In PlayScan you can send us a suggestion in your own words, from the card on the home screen or from Settings → Send a suggestion. Sending one is optional.

• What we receive: your message (up to 2,000 characters), the platform, the app version and the language, linked to your device identifier. We do not ask for your name or email address.
• Who reads it: only our team, to decide how to improve the app. Suggestions are not sent to other companies.
• How long it is kept: 12 months after you send it. Database backups may keep a copy for up to 30 more days.
• What it is not used for: training machine-learning models, advertising or sale.
• Replies: we cannot reply in the app. Please do not write personal information (such as names or phone numbers) in your message.
• Deletion: to have your suggestions deleted sooner, write to us from the app (Settings → Contact us) or at the address in Section 15.

8. Children's Privacy

Ergashev Zarifjon does not intentionally collect personal data from children under 16 and will promptly delete such information if inadvertently obtained.

9. Purposes of Personal Information Processing

The Developer may use Personal Information for:

• Communication with the User
• Marketing and promotional activities
• Statistical research and analysis
• Improving the functionality and performance of our applications
• Targeting advertising based on depersonalized data

10. User's Rights

Users have the right to:

• Be informed about data processing
• Access their personal data
• Correct inaccurate data
• Restrict processing
• Object to processing
• Withdraw consent at any time
• Request deletion of personal data
• Data portability

These rights are exercisable by contacting Ergashev Zarifjon via the contact information provided below.

11. Data Retention

Personal Information will be stored as long as necessary or permitted by law. The Developer periodically reviews data processing systems for relevance.

12. Data Protection

Ergashev Zarifjon stores and protects Personal Information against unauthorized access according to internal terms and regulations.

13. Transfer of Personal Information

The Developer may transfer Personal Information to third parties:

• With user consent
• For application functionality
• To third-party service providers
• In the event of asset transfers

14. Amendments to Privacy Policy

Ergashev Zarifjon may unilaterally amend or terminate this Policy without prior notice. Users are advised to regularly check the current version.

15. Contact

For questions, amendments, or deletion requests, contact Ergashev Zarifjon at zarifergashev4@gmail.com.