Privacy Policy

Son güncelleme: 2026-11-14
This Privacy Policy explains how Mehmet Ugur Kebir collects, uses, and protects your personal data when you use Medical Devocab (the web application at med.devocab.com and the iOS and Android apps). We comply with the EU General Data Protection Regulation (GDPR) and the German Bundesdatenschutzgesetz (BDSG).

1. Data Controller

Mehmet Ugur Kebir Peter-Dörfler-Str. 24 88161 Lindenberg im Allgäu, Germany support@devocab.com
We are a small operation and not legally required to appoint a Data Protection Officer (DPO). Direct all privacy enquiries to the email above.

2. What We Collect

As a guest

Nothing leaves your device. Study progress, settings and the free vocabulary are stored locally (see §9). Premium requires an account, so a guest never downloads the gated vocabulary.

At signup

•
Email address
•
Authentication session tokens (managed by Supabase)
•
For "Sign in with Google": the identifier, email address, name and profile picture link Google shares with us. The name and the picture link are stored with your account; the app does not use them.
•
For "Sign in with Apple": the identifier and the email address Apple shares with us (which may be a private relay address), and a token Apple issues for the app, kept only so that deleting your account can end the app’s access to your Apple ID. If the address is a private relay address, the emails we send to it go to Apple, which forwards them to your own address, and you can stop that forwarding in your Apple Account settings. If you stop or restart that forwarding, Apple tells us, and we keep which of the two applies, so that we do not send emails Apple would refuse.

During use (signed in)

•
Study progress: which terms you have studied, review history, Leitner box state, per level and language
•
Accepted synonyms (answers you have explicitly accepted as correct)
•
Error reports you submit about a term (the term, the issue type, your suggestion and note)
•
Your study-day time zone, synced to your account (the level you study, the theme and the interface language stay on the device)
•
Aggregate statistics (counts, streaks) and adaptive-learning metrics (a rolling recent-accuracy and review-load signal used to size your personalised daily plan) derived from your study activity
•
Usage timing, so that the app can show you your own study statistics: a per-day total of active study time and a running total per hour of the day, summed across your whole history. Both are counted only while you are actively studying; a pause longer than 60 seconds counts as idle and adds nothing. Hours are measured in the study-day time zone set on your account (Europe/Berlin by default). The hourly figures carry no dates: they can show that you tend to study in the evening, but not what you did on any given date. We chose this shape deliberately over a day-by-day timeline, which we do not store.
•
A daily counter of premium-vocabulary fetches per account (used solely for abuse detection and rate limiting; see §10).
•
Vocabulary watermark seed (premium users only). When a premium user first downloads a gated vocabulary level, we generate a small random seed (a number, not derived from any personal data) and store it on the account. The seed selects a per-account subset of marker entries that are merged into the premium copy, which lets us trace an unauthorised public copy back to the originating account (see Terms §6). The seed is never shared with third parties and cannot be read by any other user.

At purchase

Payment is processed entirely by the billing provider: Paddle on the web, Apple in the iOS app, Google in the Android app. We do not see, store, or have any access to your payment card data. From the provider’s notifications we receive only:
•
The provider’s customer and subscription identifiers (for the stores: the original transaction identifier)
•
Subscription status (active, cancelled, expired) and the subscription end date
Checkout on the web. When you open checkout, our server creates the Paddle transaction for your account and keeps a record of that checkout: the plan, whether it was a regular purchase or one with a promotional code, its status, the Paddle transaction, price, discount, customer and subscription IDs, and the times it was created and last changed. Paddle’s events are attached to your account through this record only. We also count the checkouts an account opens each day, to limit abuse. We pass Paddle your email address and your account ID.
Promotional codes. If you redeem a code, we store which code you redeemed, the Premium period it granted and when, plus a daily count of failed attempts (to stop codes from being guessed). A discount code is checked against Paddle’s discount list; a code that grants Premium directly involves no payment and no Paddle subscription.
Premium without a recorded source. Our security check lists the Premium accounts that no completed checkout, store subscription or redeemed promotional code explains, such as Premium we granted by hand (see Slack in §4). When the operator has checked one, that is recorded with the account ID, a fingerprint (a hash) of the subscription details checked (status, end date, provider and subscription), the time and an optional note; the record is deleted with the account.

Diagnostics

When the app crashes or hits an unexpected error, a diagnostic report is sent to our error-monitoring provider (Sentry, §4): the error message and stack trace, the app version, the operating system version and the device model. In the iOS and Android apps a report also carries other technical details of the device, such as its free memory, and the report of a crash in the app’s native code carries a random identifier of the app installation, which is not linked to your account. The reports are configured to carry no email address, no name and none of your study data or vocabulary content.

Website analytics (web only)

On the website, our hosting provider Cloudflare (§4) measures visits and page load times with Cloudflare Web Analytics, for visitors outside the European Union only: visitors from the EU are not measured, and the measurement script is not loaded for them. It sets no cookies and stores nothing on your device, does not fingerprint you, and does not follow you across other websites. It records the page address without its query string, the referring page, the browser and device type, the country derived from your IP address, and loading times, and shows them to us only as aggregate figures. The iOS and Android apps contain no such measurement.

Emails we send you

We send email to the address on your account for these purposes:
•
Sign-in emails — a link to confirm your email address when you sign up, or to reset your password when you ask for a new one.
•
Service notices — what you need to know about your account and the Service: changes to our Terms or this Privacy Policy, security incidents, outages and billing problems. They can go to any account with a confirmed email address, whether or not it asked for news and offers, so we never put advertising in them.
•
News and offers — for example new features, offers or tips. These go only to accounts that have explicitly asked for them and confirmed that request through a link we emailed (double opt-in). The app does not yet offer a way to ask for them, so at present we send none. Every such email contains a link to a page where one click unsubscribes you, and mail programs with a one-click "Unsubscribe" button can use that instead; either takes effect immediately. You can also unsubscribe by writing to us.
For service notices and news and offers we store:
•
Your choice about news and offers: whether you asked for them, confirmed or withdrew, and when.
•
For each email sent to you: the mailing it belongs to; whether it was handed to our email provider, skipped (for example because you had unsubscribed in the meantime) or failed, with the reason and when that happened; the number of attempts; and the provider’s message ID.
The operator writes each mailing and chooses its recipients in the administration tool, using the account-list fields described under "Operator access to your account data" below (for example the subscription status or the date you last studied).

We do NOT collect

•
IP addresses for tracking, profiling, or advertising. (Our hosting, CDN and backend providers — Cloudflare, Supabase, Sentry — necessarily process your IP address transiently to deliver and secure the Service, but we do not use it to track or profile you and do not store it for analytics.)
•
Browsing history outside the app
•
Precise location, contacts, biometric data
•
Marketing analytics: there are no tracking cookies and no analytics SDK; the only measurement is the cookieless website analytics described above
•
An event-level activity log. We do not store a timestamped record of individual actions, and we do not keep a day-by-day breakdown of the hours you were active. The usage timing described above exists only in the two aggregated forms listed there.
•
Whether you open our emails or click their links: they contain no tracking pixels and no tracked links. A sign-in email’s link confirms your address or lets you reset your password, and signs you in; in our other emails, the one personal link, to unsubscribe, records nothing but the withdrawal you make with it.
The statistics described in this section are shown to you, on your own Statistics page. We do not use your study times to profile you, to make automated decisions producing legal or similarly significant effects about you (GDPR Art. 22), for advertising, or for any purpose beyond operating the Service. They are never sold or shared with third parties other than the infrastructure processors in §4, who store them on our behalf.

Operator access to your account data

Besides you, the operator named in §1 can access the data held on your account, through an internal administration tool and through the backend’s administration console, both protected by multi-factor authentication. This access exists so that we can run the Service: answering your support requests, investigating a fault you report, reviewing the error reports you submit, detecting abuse of the premium vocabulary (see §10), and sending the emails described above. The administration tool is limited to what those purposes require:
•
The account list shows, for every account: email address and display name, subscription status, longest streak, number of study days, words seen, points and player level (with the language they were earned in), the dates the account last studied and was last seen, today’s premium-vocabulary fetch count, and whether an abuse alert was raised in the last 7 days. The dashboard also names the five accounts with the most premium-vocabulary fetches today.
•
An individual account’s detail view additionally shows: join date, subscription end date and whether Premium is active; learning progress (Leitner box distribution, number of reviews, accuracy); for each level and language, the current and longest streak, number of study days, last-studied date, when that level last synced, the adaptive-learning metrics described above, and points; for each language, the player level, the points still needed for the next one, and what the points were earned for (daily plans, streaks, logins, milestones, repaired days) and on how many days; total study time; abuse alerts; and the number of error reports filed.
•
Mailings. The operator selects the recipients of an email by these account-list fields only: subscription status (including Premium), longest streak, number of study days, player level, the last-studied and last-seen dates, and whether an abuse alert was raised in the last 7 days. Before sending, the operator sees the selected accounts and, for each one the mailing will skip, why (no consent to news and offers, an unconfirmed address, or an address that cannot be reached); afterwards, whom the mailing went to and whether each delivery succeeded. Sending a mailing is recorded in the audit log described below, with the kind of mailing, the number of recipients and the selection used; cancelling one records how many emails it stopped. Neither records the email’s text.
•
Aggregated across all users only, never per account: the hour-of-day study pattern and the day-by-day study-time breakdown. A figure is shown only when at least 3 accounts contributed to it. These are used to understand how the Service is used overall (for example, "most studying happens between 20:00 and 23:00"), and the administration tool offers no way to view either of them for an individual account.
•
Access to an account is logged. Whenever an administrator opens an individual account’s detail view in the administration tool, that access is recorded in an internal audit log along with the administrator’s identity and the time. Viewing the account list is not logged. As database administrator, the operator can also reach the database directly through the backend’s administration console, for maintenance, incident handling and reviewing error reports; such access is not recorded in that audit log.
Neither can read your password (we never hold it in a readable form — see §10) or your payment data (held by the billing provider — see "At purchase").

3. Legal Basis (GDPR Art. 6)

•
Contract (Art. 6(1)(b)): account creation, study features, subscription management. This includes the progress, streak, study-time and study-hour statistics: they are part of the learning product you sign up for, are shown back to you in the app, and are synced across your devices so your account behaves consistently.
•
Legitimate interest (Art. 6(1)(f)): fraud prevention, service improvement, aggregate website analytics, security logging, crash diagnostics, and the operator access described in §2. Our interest is in being able to operate, support and protect the Service. We balanced it against your interests by keeping the diagnostics free of personal and study data, restricting per-account access in the administration tool to the fields listed in §2, keeping the more revealing hour-of-day and day-by-day figures aggregate-only, logging every opening of an individual account’s detail view, and making no automated decisions about you on this basis. You may object to this processing under Art. 21 — see §7.
•
Service notices by email: Art. 6(1)(b) for information about your account and our contract with you (such as changes to our Terms); Art. 6(1)(c) where the law requires us to inform you (such as a data breach, Art. 34); and Art. 6(1)(f) for security and outage notices, in our interest in keeping you and the Service safe.
•
Consent (Art. 6(1)(a)): only where explicitly requested in-app, and for news and offers by email (together with §7(2) of the German Act against Unfair Competition, UWG). You can withdraw your consent at any time with effect for the future (Art. 7(3)); we keep a record of your consent and of its withdrawal so that we can prove and honour them (Art. 7(1), Art. 6(1)(c)).

4. Third-Party Processors

To deliver the Service we use the following third-party services, each handling your data under its own privacy terms; the data protection agreements that apply are described below the list:
•
Supabase — authentication, database and backend function hosting.
•
Paddle.com Market Limited — payment processing as Merchant of Record for web purchases.
•
Apple Inc. — in-app purchases in the iOS app, and "Sign in with Apple" (only if you choose it). Apple acts as an independent controller for the purchase and the sign-in step.
•
Google LLC — in-app purchases in the Android app (Google Play), and "Sign in with Google" (only if you choose it). Google acts as an independent controller for the purchase and the sign-in step; we receive your email address, your name, your profile picture link and the OAuth identifier (§2).
•
Cloudflare, Inc. — web application hosting, CDN and cookieless website analytics (Cloudflare Web Analytics, §2).
•
Functional Software, Inc. (Sentry) — crash and error reporting (the diagnostics described in §2).
•
Slack Technologies Limited (Slack) — receives our automated operational alerts: when abuse detection flags an account, that account’s ID and its premium-vocabulary fetch count for the day; billing alerts (for example a double purchase, a paid checkout that differs from the one our server created, a purchase paid with another billing email, or a subscription our records cannot match), with the account ID, the Paddle IDs concerned (transaction, subscription, customer, price or discount) and the kind of event and subscription status involved; the Premium accounts that no completed checkout, store subscription or redeemed promotional code explains, with their account ID, subscription status, end date, provider and subscription ID; and changes to promotional codes and how often they are redeemed (no account ID).
•
GitHub, Inc. — runs our scheduled maintenance jobs (abuse detection and the promotional-code and Premium checks); a job’s log can contain the ID of an account that abuse detection flagged, with the day and its premium-vocabulary fetch count.
•
Amazon Web Services EMEA SARL (Amazon SES) — sends our emails (§2) from its EU (Frankfurt) region: the sign-in emails, service notices and, if you asked for them, news and offers. It receives each recipient’s email address and the email itself (subject, text and, for a sign-in email, its link; for news and offers, a personal unsubscribe link). If an email to an address bounces permanently, or its recipient reports it as spam through a mail service that passes such reports on to Amazon SES (Gmail does not), Amazon SES adds that address to our account’s suppression list so that we stop emailing it.
•
Anthropic PBC and other LLM providers — used offline by us to generate or refine translations, definitions and level assignments before publication. No personal data, study progress, or user-generated content is ever sent to them; only impersonal German vocabulary terms are processed.
Our infrastructure processors provide GDPR Article 28 Data Processing Agreements (DPAs) as part of their service terms, governing their processing of personal data on our behalf. Paddle acts as our Merchant of Record; for the payment transaction it is largely an independent controller, and its DPA applies to any data processed on our behalf. Slack and GitHub receive only the limited data described above.

5. International Transfers

Where data is transferred outside the EU/EEA (e.g. via Cloudflare, Sentry, Paddle, Slack, GitHub, Amazon Web Services, Apple or Google global infrastructure), we rely on the EU Standard Contractual Clauses and our processors’ adequacy mechanisms.

6. Storage and Retention

•
Active account data is retained while your account exists.
•
When you delete your account (Settings → Delete account) or ask us to, data is removed from the live production database immediately or within 30 days, except where retention is legally required (e.g. tax-relevant records: up to 10 years per Abgabenordnung §147; payment records held by the billing provider).
•
Encrypted backup snapshots. Our database host (Supabase) maintains point-in-time-recovery backups so that the database can be restored after an incident. These backups are encrypted at rest and may continue to contain a copy of your data for the rolling backup window (typically 7–30 days) after live deletion, after which they are overwritten in the normal backup rotation. Backups are accessed only for disaster recovery; they are not used for production reads.
•
Error reports you submitted about a term (§2) are kept after your account is deleted, without any link to you or your account, because they are corrections to the dictionary. Write to us (§1) if you want one removed.
•
Administrative audit log. The record of operator accesses described in §2 is kept for 12 months and then deleted; an entry that names your account is not deleted with your account, only when it is 12 months old. It stores only which administrator opened which account, sent which mailing (with its kind, number of recipients and selection), cancelled one (with the number of emails stopped) or sent themselves a test email, and when — it holds no copy of your study data or of any email’s text.
•
Emails. The delivery record of each email sent to you as part of a mailing is deleted 90 days after its mailing finishes; when you delete your account, it loses its link to you at once. The mailing’s totals (how many emails were handed over, failed or skipped) and its text remain, with no data about you. Your choice about news and offers, with when you last confirmed or withdrew it, is kept while your account exists and deleted with it. An entry on Amazon SES’s suppression list (§4) stays until we remove it, which we do on request.
•
Diagnostic reports (§2) are retained by Sentry for 90 days.
•
Operational alerts and logs. The Slack alerts and GitHub job logs described in §4, and our server functions’ logs, can contain an account ID. They are not deleted with your account; they are kept for as long as those services retain them.
•
Website analytics measurements (§2) are kept by Cloudflare in full for 7 days and afterwards only as a reduced sample.
•
Local data on the web (localStorage and IndexedDB) is removed when you sign out or reset your progress in Settings; local data in the apps lives in an encrypted database on the device and is removed the same way, or when you uninstall the app.
•
Resetting your progress erases that level and language’s statistics immediately, on the server as well as on the device — including its study days, study time and hour-of-day totals. Study statistics are otherwise kept for as long as the account exists and are deleted with it, under the 30-day rule above.

7. Your Rights (GDPR Art. 15–22)

You have the right to:
•
Access your personal data (Art. 15)
•
Rectify inaccurate data (Art. 16)
•
Erase your data ("right to be forgotten", Art. 17) — the in-app account deletion does this without contacting us
•
Restrict processing (Art. 18)
•
Data portability — receive your data in a structured, machine-readable format (Art. 20); email us and we send you a JSON export of your account
•
Object to processing (Art. 21)
•
Withdraw consent at any time
To exercise any of these rights, email support@devocab.com. We respond within 30 days as required by law.

8. Supervisory Authority

You have the right to lodge a complaint with a supervisory authority. The competent authority for us is:
Bayerisches Landesamt für Datenschutzaufsicht (BayLDA) Promenade 18, 91522 Ansbach, Germany www.lda.bayern.de

9. Cookies and Local Storage

On the web we use localStorage and IndexedDB (browser APIs, not cookies) for:
•
Authentication session tokens (necessary)
•
Application state and user settings (necessary)
•
The pending sync queue, the study progress cache and the downloaded vocabulary (necessary)
In the apps the same data lives in an encrypted database and key-value store on the device, keyed from the device keychain. We do not use marketing or analytics cookies. Supabase Auth may set a session cookie for its own auth flow on its domain (not ours). No third-party cookies are set by our site.

10. Security

We apply technical and organisational measures appropriate to the risk:
•
HTTPS / TLS 1.2+ for all data in transit.
•
Database access restricted by Postgres Row-Level Security: each user can read and write only their own data.
•
Subscription status fields are written only by our server-side functions — from the billing providers’ signed notifications (Paddle HMAC, Apple App Store Server Notifications, Google Play real-time developer notifications), or when you redeem a promotional code — never by the app itself; clients cannot modify them.
•
Authentication uses industry-standard JWT bearer tokens. Passwords (where you use the email/password sign-in option) are never seen or stored by us in plaintext; they are hashed and salted by Supabase Auth using bcrypt before storage. We have no access to your password. Sign in with Apple or Google does not involve a password held by us at all.
•
Service-role database credentials and third-party API keys are stored only in server-side environment variables and are never sent to the app or the browser.
•
On the phone, local data is stored in an encrypted database whose key lives in the device keychain.
•
Premium vocabulary delivery is rate-limited per account; abnormal access patterns trigger an alert to the operator (see Slack in §4).
No security measure is perfect; despite our efforts we cannot guarantee that data transmission or storage will be free from unauthorised access.

11. Data Breach Notification

In the event of a personal-data breach that is likely to result in a risk to your rights and freedoms, we will notify the competent supervisory authority (BayLDA, see §8) within 72 hours of becoming aware of the breach, in accordance with Article 33 GDPR, and notify affected users without undue delay where the breach is likely to result in a high risk to their rights and freedoms, in accordance with Article 34 GDPR — via the email address on the affected account and, where appropriate, an in-app banner. The notification will describe the nature of the breach, the categories and approximate number of users and records concerned, the likely consequences, and the measures taken or proposed to address it.

12. Children

The Service is not directed to children under 16. We do not knowingly collect data from children. If you become aware that a child has registered, please contact us and the account will be deleted.

13. Changes

Material changes to this Privacy Policy will be communicated by email or in-app banner. Continued use after the effective date constitutes acceptance.
© 2026 Mehmet Ugur Kebir · devocab.com