Privacy Policy
The short version: the iPhone app keeps your learning on the device and has no account. The website study app saves your account, progress and purchases so you can return to them. Optional website analytics starts only if you choose Accept all. You can decline or change your choice without losing access or saved progress.
Last updated: 13 November 2026
1. Who is responsible
The controller responsible for data processing under the EU General Data Protection Regulation (GDPR) is:
Jason HolwegViktoriastraße 3a
24937 Flensburg, Germany
Email: hello@imatlearn.com
We have not appointed a Data Protection Officer, as we are not legally required to do so.
2. The Mito app
2.1 What stays on your device
The learning core of Mito runs on your iPhone: lessons, practice, tests, your study plan, your progress and your notes are stored in a database on your device and are never transmitted to us. In particular:
- No account, no registration, no login. We never ask for your name, email address, or any other identifier.
- No analytics, no tracking, no advertising. The app contains no analytics SDK, no advertising framework, no third-party tracking library, and no advertising identifier (IDFA).
- No crash or usage reporting to us.
- No transmission of your learning data. Your answers, progress, mastery estimates, notes and readiness estimates never leave the device.
We therefore hold no profile of you and cannot identify you. Deleting the app removes this data from your device. If your device backs up to iCloud, Apple's backup may include this app data; that backup is governed by Apple's Privacy Policy and is encrypted under your Apple Account.
Three feature groups do communicate with the internet, and only these: Mito (2.2), content reports (2.3), and purchases and codes (2.4). Everything else works offline.
2.2 Mito, the AI assistant
In the web study area: if you use Mito, your question, recent replies and the current lesson step are sent through our server to the AI provider described below. We store the conversation and request status with your web account so you can reopen it and retry without counting a successful request twice. The web allowance is five free messages per account; the IMAT Pass has an abuse limit of 80 messages per UTC day. Web and iOS allowances are currently separate. Deleting your web account removes its conversations. Lesson answers and help actions are stored with their content version; supported lesson answers are distinguished from independent recall when calculating learning progress. Typed lesson notes are saved with your lesson checkpoint and are not sent to Mito unless you include them in a question.
Mito answers your questions about the study material. To do that, what you type has to leave your phone — an assistant that never receives your question cannot answer it.
What is sent when you ask Mito something:
- The messages of that conversation — your questions and Mito's previous replies.
- The title of the concept you opened the chat from, as context (for example “Bond polarity and partial charges”).
- A fixed instruction text that defines the role of Mito or the friend you are talking to, identical for every user.
- If you tap Lysosome after a wrong answer: that question, your chosen answer and the official explanation.
- Only if you allow it, when you chat with a friend: a short study summary (days until your exam, how much of each section you have covered, completed mocks, today's plan and the topics of recent mistakes). You can turn this off in the Mito settings.
- A random device ID generated by the app. It contains no hardware, Apple Account or advertising identifier and is used only to enforce the free-message allowance and abuse limits.
What is never sent: your individual answers (other than the one above), your test results, your notes, your name, email address or Apple Account.
Where it goes: the request travels to a proxy we operate on Cloudflare Workers, which forwards it to Anthropic PBC (San Francisco, USA), whose model generates the answer. The proxy exists so that no API key has to sit inside the app. Under Anthropic's commercial API terms, inputs are not used to train their models; their handling is described in Anthropic's Privacy Policy.
- Legal basis: Art. 6(1)(b) GDPR — performing the service you asked for by opening the chat.
- Recipients: Cloudflare, Inc. and Anthropic PBC as our processors.
- Transfer outside the EU: both process data in the USA, safeguarded by EU Standard Contractual Clauses.
- Allowances and rate limiting: to enforce the five-message free allowance, the proxy stores only a count under the random device ID. Short-term device and IP abuse counters are deleted within 48 hours. The counters hold no message content (Art. 6(1)(b) and (f) GDPR — providing the allowance and our legitimate interest in a functioning, affordable service).
- Conversation history: your chats are stored on your device so you can reopen them. Deleting a chat, or the app, removes them.
Please do not type personal information into the chat — names, contact details, health information. Mito needs none of it to explain a concept, and anything you send is transmitted as described above.
Mito is optional. If you never open the chat, nothing is ever sent.
2.3 Content reports
If you find a problem in a question, explanation or concept card, you can deliberately send us a report. The app sends the content source ID, the reason you choose, your optional report note, the reported question or card text as context, a random report ID, the app and content-pack version, and your language setting. It does not send your answer, progress, test result or account information. The request uses the same random app-generated device ID only for a 48-hour abuse counter; the report stored in the inbox contains no device ID.
Reports are stored in a database operated by us on Cloudflare so we can review and correct the affected content. They are not forwarded to Anthropic or another AI provider. The legal basis is Art. 6(1)(b) GDPR: processing the report you chose to submit. Open reports are retained until they have been reviewed; closed reports are automatically deleted after six months. Please do not include names, contact details or other personal information in the optional note.
2.4 Purchases and codes
The IMAT Pass is purchased through Apple in the App Store. Apple is the merchant of record: you enter your payment details with Apple, not with us. We never see or store your payment data, and we receive only anonymous, aggregated sales reports. Apple's handling is governed by Apple's Privacy Policy. Restoring a purchase works through your Apple Account. The iPhone app has no account of our own. The website study app does; that is described in section 3 and does not apply to the iPhone app.
If you redeem a discount or gift code, the app sends only that code to our Cloudflare Worker, which checks whether it is valid and whether a gift code has already been used. No personal data is transmitted with it (Art. 6(1)(b) GDPR).
2.5 Notifications
The app can send study reminders. These are local notifications, scheduled by the app on your device — no push server is involved and no device token is transmitted to us. Notifications are optional: you decide whether to allow them when asked, and you can change your mind at any time in iOS Settings.
3. This website and the web study app
This section covers imatlearn.com and app.imatlearn.com. It does not change section 2. The promise that learning stays on the device applies to the iPhone app only.
3.1 Pages that do not need an account
The marketing pages, the syllabus, the guides and the free diagnostic do not require an account. They embed no third-party fonts, videos or social widgets. Optional first-party analytics is described in section 3.12; separately optional ChatGPT advertising measurement is described in section 3.14. We remember your privacy choice whether you accept or decline; the analytics identifier is set only after you accept.
3.2 The diagnostic
The free diagnostic on this website runs entirely in your browser. Your answers are never transmitted to us — there is no endpoint to send them to. The questions are delivered as a static file from this same website, exactly like an image or a stylesheet, and everything after that happens on your device.
So that reloading the page does not throw away a test you are halfway through, your answers are kept in your browser's session storage while the test is running. This is strictly necessary to provide the feature you asked for (§ 25(2) no. 2 TDDDG), which is why it needs no consent banner. It is not a cookie, we never receive it, and we do not use it to recognise you: the entry belongs to a single browser tab and is normally discarded when you close that tab. Some browsers offer to restore a session after a crash or a restart, in which case the entry can survive that long — it stays on your device either way. You can clear it yourself at any time through your browser's settings, or by choosing “Take it again” at the end of the test.
We do not receive which option you chose. If you want to keep the full result, take a screenshot before you close the tab — we cannot restore the questions or your answers, and we could not identify them as yours even if you asked.
Before the test starts, you can choose to share statistics. That is optional, and skipping it does not change the test. If you confirm, we store a random id for that attempt, when it started and finished, how many answers were correct, wrong or skipped (including a count per subject), and whether you then opened the App Store page or the web study app. We do not store your answers, your account, or your name with that record. You can stop sharing for the rest of the session from the result screen. Choosing “Only essential” in Privacy preferences also stops diagnostic sharing and skips its optional sharing prompt. Accepting website analytics does not itself enable these separately consented diagnostic statistics. Shared counts are kept for up to a year and then deleted. The legal basis is your consent (Art. 6(1)(a) GDPR). You can withdraw it for future tests by choosing Skip the next time; counts already stored stay in the aggregate and are not tied to an account we could use to pick yours out.
3.3 Server logs
The website is hosted by Cloudflare, Inc. (101 Townsend St., San Francisco, CA 94107, USA). Like every web server, Cloudflare's infrastructure processes technical connection data when you open a page — typically your IP address, the requested file, date and time, referrer, and browser and operating system version. This is technically necessary to deliver the website and to protect it from attacks.
- Legal basis: Art. 6(1)(f) GDPR — our legitimate interest in operating a functioning, secure website.
- Recipient: Cloudflare acts as our processor under a data processing agreement (Art. 28 GDPR).
- Transfer outside the EU: Cloudflare may process data in the USA. The transfer is safeguarded by the EU Standard Contractual Clauses and Cloudflare's certification under the EU–US Data Privacy Framework.
- Retention: Cloudflare retains such logs only briefly, for security and troubleshooting purposes.
3.4 Contacting us by email
If you write to us, we process your email address and the content of your message solely to answer you (Art. 6(1)(b) and (f) GDPR). We delete such correspondence once it is no longer needed and no retention obligation applies.
3.5 The web study account
Lessons, practice, mocks and digital-book purchases on the website use an account. You create it with an email and a password, or with Google. The same login works in another browser and on another device. That is the point of it: a purchase and your progress are not trapped in one browser.
Creating the account is not optional if you want to use the web study app. The registration form asks you to agree to the Terms of Use and acknowledge this policy. We save the account and learning progress needed to provide the study app. We do not pre-tick that box. Continuing with Google uses the same agreement. Optional analytics has its own choice and is never a condition of registration or access.
What we store for a web account:
- Your email address, and, if Google sends one, your name. A password account stays closed until you open the confirmation link we email you. Google has already confirmed the address, so that path does not send a second link.
- A password hash and salt, if you chose a password. We do not store the password itself. A Google-only account has no password.
- If you use Google: Google's account identifier (the
subclaim), so we can recognise the same Google account next time. - The time you agreed, the time the account was created, and the time the email address was confirmed.
- A confirmation token, stored only as a hash, until you use the link or for 24 hours, whichever comes first.
- A session token, stored on our side only as a hash, in an HttpOnly cookie on your browser. It lets you stay signed in. It is not an advertising identifier.
- Your study-plan preferences: your chosen target date, daily study time, study days and time zone. Saved daily routes, lesson completions and recall answers help us resume your work and schedule practice. Earlier progress is retained when the learning model is upgraded. Daily progress snapshots store the model version, coverage, understanding, retention, timed accuracy and any estimated score range. Optional confidence ratings help explain calibration. Web and iOS activity are not synced yet.
- Your learning progress: which concepts you marked as read, answers, time taken, unfinished sittings, finished sittings and the repair that follows a sitting.
- Past-paper simulator attempts: shuffled question and answer order, selected answers, start and finish times and results. These are tied to your web account for resuming and review, kept separately from readiness estimates, and removed when you delete the account.
- Operational records needed to run the account, deliver lessons and practice, resume your work, prevent duplicate submissions, and provide purchases. Saved learning answers and progress continue to work when optional analytics is off.
- What you bought: the product (the IMAT Pass, a mock pack, a single mock, or a digital book), the amount, the time, and the Stripe checkout reference.
- For digital books: the order and payment references; your request for immediate delivery and withdrawal acknowledgment, including the accepted wording, its version and the acceptance time; and the purchase confirmation email address, text and sending status. Refund or dispute records contain the payment reference, reason and time.
We do not load the ChatGPT advertising SDK in the signed-in study app. The required activity above is stored in our own database to run your account and learning features. Optional product analysis is separate and requires the choice described in section 3.12. We do not sell your data, and we do not use it to advertise to you on other sites.
3.6 Why, and the legal basis
- Running the account, syncing progress, and unlocking what you bought — Art. 6(1)(b) GDPR. This data is required to provide the web study app you signed up for.
- Optional website analytics to improve the pages and learning material — Art. 6(1)(a) GDPR and § 25(1) TDDDG. Collection begins only after a separate, freely given choice in Privacy preferences. Declining or withdrawing does not affect account access, lessons or saved progress. This does not apply to the iPhone app.
- Keeping a purchase record — Art. 6(1)(c) GDPR, where tax and accounting law requires us to retain it, and Art. 6(1)(b) for providing the access you paid for.
- Confirming the email address — Art. 6(1)(b) GDPR. The link is how we know the address belongs to the person creating the account.
- Security of the account — Art. 6(1)(f) GDPR, our legitimate interest in preventing abuse of sign-in and checkout. Repeated sign-in and registration attempts are counted for about 15 minutes. The counter is stored under a hash of the connection address and the kind of attempt, not next to your account, and the address itself is not kept in that counter. We do not add your IP address to the learning record. Connection logs are the ones in section 3.3.
3.7 Google sign-in
If you choose “Continue with Google”, your browser goes to Google and comes back to us. We receive the email address Google verifies, your name if Google provides it, and Google's account identifier. We use them only to create or open your account. We do not receive your Google password, and we do not post to your Google account.
Google is a separate controller for the sign-in page. Google's Privacy Policy describes that step. The legal basis for our use of what Google sends back is Art. 6(1)(b) GDPR.
3.8 Payments
Web purchases are taken by Stripe. You enter the card on Stripe's page. We do not receive the card number. Stripe tells us whether the payment succeeded, the amount, the product, and an email address for the receipt. We store the product, the amount, the time and Stripe's checkout reference against your account, and we unlock that product on the account.
Depending on the product, a web purchase unlocks website study features or an online browser edition of a digital book. It does not unlock the iPhone app. The iPhone app remains an Apple in-app purchase, described in section 2.4. Stripe may set cookies on Stripe's own pages. Stripe's handling is described in Stripe's Privacy Policy.
For a digital book, we send the purchase confirmation to the email address on your account. Browser reading becomes available after payment is verified and that confirmation has been sent. We use the purchase and payment records to provide access and handle refunds or disputes. We store your last-read page so you can resume reading. Book pages carry a personal licence mark linked to your purchase to discourage unauthorised sharing. Reading progress and access checks are necessary to provide the book and do not depend on optional analytics consent.
3.9 The session cookie
After you sign in we set one cookie, mito_session. It is
HttpOnly, SameSite=Lax, and on the production site it is Secure and scoped
to .imatlearn.com so imatlearn.com and app.imatlearn.com share
the login. It lasts 30 days, or until you sign out or delete the account.
The cookie is strictly necessary to keep you signed in to the account you asked for (§ 25(2) no. 2 TDDDG). It is not used to track you across other websites. Optional advertising cookies are described separately in section 3.14. The Privacy preferences card controls optional analytics, not the session cookie.
3.10 How long we keep web-app data
- Account, progress and required operational records — for as long as the account exists. Optional analytics has the shorter retention in section 3.12.
- Session — 30 days from sign-in, or until you sign out. Deleting the account deletes its sessions. No session is created until the email is confirmed.
- Confirmation link — 24 hours, or until you use it. The sign-in counter is discarded after about 15 minutes.
- After you delete the account — we delete the login, the progress, the associated optional analytics and the unlocks. Account and browser-book access end, and saved reading progress is deleted. A purchase row (email, product, amount, date, Stripe reference) is kept for as long as German accounting rules require, which can be up to 10 years, and is then deleted. That row does not restore the account. Book order, payment, delivery-acknowledgment and confirmation records are unlinked from the deleted account and retained with the purchase record; refund or dispute records are also retained. They do not restore browser-book access.
- You can also ask by email if you cannot sign in. Write to hello@imatlearn.com from the address on the account.
3.11 Who receives web-app data, and transfers
- Cloudflare hosts the site and the database (D1). Cloudflare is our processor (Art. 28 GDPR). Details and the transfer safeguard are in section 3.3.
- Google receives the sign-in request only if you choose Google. See section 3.7. Google may process that request in the USA under its own safeguards, including the EU–US Data Privacy Framework where it applies.
- Stripe receives the checkout. Stripe may process payment data in the USA. The transfer is safeguarded by the EU Standard Contractual Clauses and, where applicable, Stripe's certification under the EU–US Data Privacy Framework.
- Resend sends account confirmation emails, digital-book purchase confirmations, and the notice if someone tries to register an address that already has an account. Resend receives that address and the message. Resend's privacy policy describes that processing. Resend may process it in the USA; the transfer is safeguarded by the EU Standard Contractual Clauses.
We do not pass your learning progress to an advertising network, a data broker, or the iPhone app.
3.12 Optional website analytics and your choice
We ask whether you want to help us improve Mito using first-party analytics. “Accept all” and “Only essential” are both available on the card. “Accept all” turns on website analytics and, on public marketing pages, ChatGPT ad measurement (section 3.14). Closing the card, browsing, signing in or creating an account does not count as consent. Analytics stays off until you accept. You can reopen here, in the site footer, or from your web account, and choose “Only essential” to withdraw without deleting your account.
If you accept, we collect page categories visited, resource or App Store clicks, and selected learning interactions so we can identify confusing lessons and useful resources. Website events contain a random event ID, time, a limited page path and a predefined action or resource category. On the IMAT results page, we also measure resource clicks and subsequent account creation in the same consenting browser, within 30 days. We derive only a broad arrival category (recognised search engine, internal link, referral, or direct/unknown); we do not store referrer URLs or search terms. Account-creation measurement uses the actual server record and can be associated with your account so we can distinguish created and confirmed accounts. No analytics is collected for these purposes before acceptance. For signed-in learning activity, optional analytics can be associated with your account and the learning record needed to understand that interaction. We do not put typed answers, notes, chat content, email addresses, URL query strings or raw URL fragments in these first-party analytics events. They are separate from the optional OpenAI advertising measurement in section 3.14. We do not use fingerprinting or session replay for these analytics.
- Remembering your choice:
mito_privacyremembers the saved choice andmito_privacy_intentremembers your latest choice so a delayed response cannot restart analytics after refusal. Both contain only the policy version and whether you accepted or declined, and expire after up to 180 days. These preference cookies are necessary to respect your choice (§ 25(2) no. 2 TDDDG). Matching choice, expiry and revision values in local storage prevent an older response in another tab from undoing your refusal. They expire after 180 days; expired local values are removed on your next visit. This essential preference state contains no analytics identifier, activity or usage counters; declining does not create an analytics identifier. - Optional analytics cookie:
mito_analyticsholds a random identifier only after acceptance. It lasts up to 180 days and is HttpOnly, with Secure and SameSite protection in production. Withdrawing removes it. It is not used to track you on other websites. - Retention: raw optional analytics is kept for up to 90 days. Your choice is remembered for up to 180 days, after which we ask again.
- Withdrawal: stops new optional collection and deletes raw website analytics linked to this browser’s analytics identifier. It does not affect the lawfulness of earlier collection. You may also request deletion of personal data already collected by contacting us. Account and progress storage needed to provide the service continues.
- Provider: we operate the analytics ourselves using Cloudflare hosting and D1, with the safeguards described in section 3.3.
3.13 Reporting a web lesson
When you choose the flag in a lesson and send a report, we save the lesson and concept identifiers, content version, step, selected reason, optional comment, submission time and your account identifier. This lets us find and correct the material you reported. Reports work without optional analytics. Please avoid personal information in the comment.
Only authenticated administrators can read and manage these reports. We process them to respond to your request and maintain accurate learning material (Art. 6(1)(b) and (f) GDPR). Comments and account links are removed after a report has been closed for 180 days, or when you delete the account. We retain the non-personal content reference, report reason, status and repair/release history so that a resolved error does not lose its audit trail. Web reports are stored separately from iPhone reports in section 2.3.
3.14 Optional ChatGPT ad measurement
We advertise Mito in ChatGPT. On public marketing pages, you can enable “Allow ChatGPT ad measurement” in Privacy preferences, either with its own checkbox or by choosing “Accept all”, which turns on website analytics and ad measurement together. This setting is off by default. The checkbox saves immediately when you change it, so you can keep analytics and switch ad measurement off again at any time. Registering or purchasing does not enable it. “Only essential” turns off both optional categories. An analytics permission given before “Accept all” existed is not treated as advertising permission.
After this opt-in, we load OpenAI’s Measurement Pixel and send a page-view event so OpenAI can attribute visits to our ChatGPT ads. Requests go to OpenAI at bzrcdn.openai.com and bzr.openai.com. OpenAI receives normal connection information such as your IP address and browser information, along with the event time, site origin and advertising attribution/browser references. Our event payload does not include account IDs, names, email addresses, learning answers or payment details. This integration does not load on account, lesson, diagnostic, book-reader or admin pages. We send the event with OpenAI’s opt-out flag for future user-level personalization.
- Choice: the essential
mito_adscookie remembers accepted or declined for up to 180 days. The local-storage entrymito.ads.preference.v1only signals a preference change to other tabs; it contains no visitor identifier or browsing activity. - Advertising cookies after opt-in:
__opprefstores an OpenAI ad-click reference for up to 30 days;__obrefstores a random browser reference for up to 365 days. These are cookie lifetimes, not a statement about OpenAI’s server-side retention. - Withdrawal: switch the option off in Privacy preferences, or choose “Only essential”. This stops future events and removes these advertising cookies. It does not erase events already received by OpenAI. Your account and purchases continue to work.
- Provider information: see OpenAI’s Privacy Policy for its data handling, international transfers and your rights, and OpenAI’s Measurement Pixel documentation for cookie and consent behaviour.
4. Your rights
Under the GDPR you have the right to access (Art. 15), rectification (Art. 16), erasure (Art. 17), restriction of processing (Art. 18), data portability (Art. 20), and to object to processing based on legitimate interests (Art. 21). Where we rely on consent, you may withdraw it at any time. Use Privacy preferences to turn off optional website analytics while keeping your account and progress. To delete the account, use Delete account in the web study app. To exercise other rights, or ask about data already collected, write to hello@imatlearn.com.
For the iPhone app, we still hold no learning profile. Deleting the app removes what it stored on the device. The only potentially personal data from the iPhone app on our side is text you deliberately submit through Mito or a content report, plus the rate-limiting counters in sections 2.2 and 2.3. The random device ID is not connected to a name, to a web account, to an advertising profile, or to data from other services.
For the web study app, we can identify the account by the email address. A copy of the progress stored for that account is available on request. Deletion is described in section 3.10. A purchase record kept for a legal retention duty is the exception to erasure for that period (Art. 17(3)(b) GDPR).
You also have the right to lodge a complaint with a supervisory authority (Art. 77 GDPR), for example the data protection authority of your country or of our registered office.
5. Users outside the EU
We apply the same standard worldwide. For residents of California: we do not sell or share personal information within the meaning of the CCPA/CPRA, and we do not use it for cross-context behavioural advertising. Web-app activity is used to operate the account and to improve the product, as described in section 3. iPhone-app data is limited to sections 2.2 through 2.4.
6. Children
The service is aimed at people preparing for a university admission test, who are typically 17 or older. It is not directed at children under 13. The web study account asks you to confirm that you are at least 16. If you are under 16, do not create one. If a younger child's data has been stored, write to us and we will delete the account.
If you are a minor under the law of your country, please read the Terms of Use regarding parental consent for purchases, and take particular care not to type personal details into the Mito chat or an optional report note.
7. Changes to this policy
We will update this policy when the product changes. The version in force is always the one published here, with the date shown at the top of this page. The web study app and the iPhone app are described separately on purpose. A change to one does not silently change the other.
8. Independence
Mito - Learn IMAT is an independent exam preparation platform and is not affiliated with, endorsed by, or associated with the Italian Ministry of University and Research (MUR), CINECA, Cambridge University Press & Assessment, or any participating university.