Retrieva Back to the site

Privacy statement

Version: 2026-11-16.1

Retrieva helps you study with homework, tests and practice questions. To do that we need some data about you. This statement explains which data, why, how long we keep it and what you can do yourself. We never sell data and we do not track you across the internet.

1. Who is responsible?

The controller for your personal data is Codea Solutions. Retrieva is a service of Codea Solutions. Registered with the Dutch Chamber of Commerce (KvK) under number 42180669. VAT number: NL005561341B04. Business address: Prinses Beatrixstraat 5, 5953 LL Reuver. All company details are also listed under company details.

Questions about privacy? Email info@retrieva.eu. We have not appointed a data protection officer because it is not mandatory for our activities; that email address is your point of contact.

2. Which data do we process, for what purpose and on what legal basis?

The table below shows, per type of data, why we use it and on which legal basis under the GDPR this happens.

DataPurposeLegal basis
Account: name, email address and password (we store the password only as a hash). If you log in with Google: your name, your email address and your Google ID (a fixed number Google uses for your account). We ask Google for nothing more. We never see or receive your Google password: you only type it at Google itself.Creating an account, logging in and letting you find your data again. With Google also: recognising you when you come back, and linking your sign-in to an existing account with the same email address confirmed by Google, so you do not end up with a second account.Performance of the contract (Art. 6(1)(b) GDPR)
What you create and upload: subjects, questions, answers, notes, flashcards, tests, planner entries, results, and photos or files of homework and tests.Providing the service: making explanations and practice questions, showing your progress and having you repeat things at the right moment.Performance of the contract
Settings: language, theme, daily goal and reminders.Making the app work the way you want.Performance of the contract
Usage data: the days and hours you were active, how many AI actions you used per day and roughly what that AI cost us per day. If an administrator switches off the AI on your account because of misuse, we keep that too.Showing your progress, showing how much you have left (Settings → Your usage), safeguarding the limits of Fair use and keeping AI costs manageable.Performance of the contract and our legitimate interest (Art. 6(1)(f))
Payments: chosen subscription, amount, date, status, the Mollie reference and the identifier of your customer record and mandate at Mollie (for automatically renewing a monthly subscription). We do not see your bank or card details; they go straight to Mollie. Only if you pay by bank transfer yourself (for example for an invoice), or if we refund something by bank transfer, do we see your name and account number as they appear on our bank statement.Processing payments, automatically renewing a monthly subscription, emailing you beforehand about a renewal or the end of your subscription, sending invoices and keeping our accounts.Performance of the contract and legal obligation (Art. 6(1)(c): tax record-keeping)
Agreement: when you agreed to the terms and privacy statement, the version, and your confirmation of your age. For a payment also your express consent to start immediately.Proving that you agreed, and that you expressly waived the right of withdrawal.Legal obligation and legitimate interest
Origin at sign-up: the channel you came from (for example “tiktok” or “meta”), the campaign and the ad from the link you clicked (this travels in the link to the app; we store nothing on your device for it), the language page, a school or promo code if you used one, your country (derived from your browser’s language setting, so not from your IP address) and whether you signed up with email or with Google. If a Meta ad adds a Meta tracking number to the address (fbclid), we remove it straight away and do not store it. We do not use the Meta Pixel or the Conversions API, set no Meta cookies and send nothing to Meta. From Meta we only fetch the cost and reach of our own ads, without any data about people.Finding out whether our promotion works: per campaign and per ad, what it cost next to how many sign-ups and payments it brought. One line per account; no profile and no tracking pixels.Legitimate interest (Art. 6(1)(f)). You can object.
Counting site visitors: without cookies. We make a daily hash of address and browser with a secret that changes every day and is never stored. The next day the hash can no longer be traced back. We count the channel, campaign and ad from the link along with it.Knowing how many people visit the site and click through, also per ad.Legitimate interest. No personal data remains.
“Isn’t this right?”: if you report an explanation, practice question, test question or flashcard, we keep what was on screen (the question, the answer and the explanation), the subject, the AI model used, the time, your account and your comment if you gave one (at most 500 characters).Finding and fixing mistakes in explanations and questions.Legitimate interest (Art. 6(1)(f)): the quality of the service. Reporting is voluntary.
Reason for cancelling: if you choose a reason when you cancel, we keep that choice, your own comment for “something else” (at most 500 characters), your plan and the date. The question is optional.Understanding why people stop and improving the service. In the admin panel we only see numbers per month and the individual comments.Consent (Art. 6(1)(a)): you fill it in voluntarily.
Reason for deleting your account: if you choose a reason when you delete your account, we keep that choice, your own comment for “something else” (at most 500 characters), whether you had a trial or a subscription, and the date. Without your name, email address or any other link to you: your account is gone afterwards. The question is optional.Understanding why people stop and improving the service. In the admin panel we only see numbers per month and the individual comments.Consent (Art. 6(1)(a)): you fill it in voluntarily.
Security: we keep your IP address only briefly in working memory to limit too many login attempts. It does not enter the database.Preventing abuse and attacks.Legitimate interest
Confirming your email address and new accounts: when you sign up with email we send a link to confirm your address; we only keep a hash of that link (valid for 7 days) and the moment you confirmed. To see how many accounts are created per day from one internet connection, we keep an encrypted code of your IP address (an HMAC that cannot be turned back into the address), the day and your account. We do not keep the IP address itself.Preventing misuse of the free trial and the AI with fake addresses or many accounts.Legitimate interest (Art. 6(1)(f))
Content check: what you type or upload for an AI feature (questions, topics, pasted text, photos and files) is automatically checked for inappropriate content via OpenAI’s moderation service (a processor we already use), without your name or email address. We only keep the category and the time of a hit, never the content itself. An inappropriate upload is deleted immediately and does not end up in backups. If it is about self-harm, you get a message with help and we only count the number, without your account.Protecting children and young people, keeping the AI for schoolwork only and preventing misuse.Legitimate interest (Art. 6(1)(f))
Email: a welcome email after signing up explaining how to start, a reminder email when something is ready to review (on by default; you can switch it off at any time in Settings or with the link at the bottom of every email), notices about your trial and your subscription (such as 2 days before the trial ends) and, only if you set it up, a weekly summary to the address of a parent or carer. That parent or carer first gets one email to confirm it themselves; without that confirmation we send nothing.Helping you repeat in time and informing you about your subscription.Performance of the contract for reminders about your own study work (reviewing in time is what Retrieva is for) and for subscription notices; consent (Art. 6(1)(a)) for the weekly summary.
Family: with a Family subscription, the email address the payer invites, and who is in which family. The payer only sees the name and email address of family members.Inviting family members and giving them access.Performance of the contract (Art. 6(1)(b)).
News and messages: now and then an email about new features or a Retrieva promotion, only to people who have or had a subscription. You can switch it off right when you sign up, every email has an unsubscribe link, and you can switch it off in Settings.Telling you what is new in the service you already use.Legitimate interest (Art. 6(1)(f)) and Art. 11.7(3) Dutch Telecommunications Act (own customers, own service).
Messages to us: what you email or write to us.Answering your question.Legitimate interest and performance of the contract

We do not use your data for advertising, we do not sell it and we do not share it with data brokers.

3. How the AI works with your data

Retrieva creates explanations, practice questions and summaries with artificial intelligence from Anthropic. When you offer a photo, file or text, we send its content (and settings such as language and level) to Anthropic to have an answer made. We do not send your name or email address along.

According to Anthropic’s terms for commercial API users, data submitted through the API is not used to train their models. Anthropic may retain it temporarily for safety and abuse prevention. Please read their current terms for the exact periods.

What we ask of you: do not photograph or upload health data, names or data of classmates or other people, or passwords. If that is on a sheet, cover it up. What you upload we process as described above.

No automated decisions with effects: the AI can assess an answer to a practice question, but that is feedback for yourself. No decisions with legal or similarly significant effects are taken about you. You can always change the assessment yourself. AI can make mistakes; always check important answers against your own course material.

4. Children and young people

Retrieva is intended for pupils and students. You must be 16 or older to create an account yourself. If you are younger, permission from a parent or carer is required. When creating an account we ask you to confirm that you are 16 or older, or that a parent or carer gives permission. We do not verify this ourselves; the parent or carer is responsible for that permission. Reminder emails are only about your own study work, are on by default and can be switched off with one click. The weekly summary only ever goes to a parent or carer who confirms it themselves by email. We comply with the GDPR, including the extra protection it gives children, and with our AI provider’s guidelines for services that are also used by minors: everything you give the AI is automatically checked for inappropriate content, the AI sticks to schoolwork and is set up for users from age 12, we monitor for misuse, and it is always clear that you are working with AI and not with a person.

Are you a parent or carer and do you want to see, correct or delete your child’s data? Email info@retrieva.eu. We will then check that you are the parent or carer. Even if a parent receives a weekly summary, the child’s answers and notes stay private: the summary contains the time practised (also per subject), the number of questions and the percentage correct, the topics that are still difficult (with the percentage correct) and the tests planned in the next two weeks. No questions, answers or notes.

5. Who do we share your data with?

We use companies that process data on our behalf (processors). With Anthropic, Google and OpenAI their data processing agreements apply, including terms on security and confidentiality. We use Replit under its terms of service and privacy policy; Replit only offers a separate data processing agreement on more expensive plans. They are:

PartyPurposeCountry
Anthropic, PBCArtificial intelligence: explanations, practice questions, summaries and assessment of answers.United States
Replit, Inc.Hosting of the app, the database, uploaded files and backups.European Union (Replit, Inc. itself is based in the United States)
Google (Google Ireland Ltd. / Google LLC), via Google WorkspaceGoogle Drive of our business Workspace account, where we keep the backups encrypted (AES-256-GCM), away from the server so they are not lost together with it, and the PDFs of invoices; and sending email (reminders, subscription notices, invoices). The Google Workspace data processing agreement applies.Ireland / United States
OpenAI Ireland Ltd. (and OpenAI, L.L.C.)Automatic check for inappropriate content: what you type or upload for an AI feature (text, photos, files) first goes through OpenAI’s moderation service, without your name or email address. And only if photos for questions are switched on: a description of the desired photo that our AI writes for a question, to create an image for it.Ireland / United States
Mollie B.V.Payments via iDEAL and card. Mollie is also responsible for payment data itself and has its own privacy statement.Netherlands
Google Ireland Ltd.Only if you log in with Google. Google decides for itself what it does with your login activity.Ireland (may involve transfers outside the EEA)

We otherwise only provide data to authorities or other parties if we are legally required to do so.

6. Transfers outside the European Economic Area

Anthropic, Google and OpenAI process data (also) in the United States, so outside the EEA. Replit stores the app, the database and the files in the European Union, but is a US company. Per party, this is the basis for that transfer:

  • Anthropic: the European Commission’s standard contractual clauses, included in Anthropic’s data processing agreement.
  • Replit: the data is stored on Replit servers in the European Union. Replit is a US company and can therefore also access the data from the United States; its terms of service and privacy policy apply to that. For our plan there is no separate data processing agreement with standard contractual clauses.
  • Google: the EU-US Data Privacy Framework (Google LLC is certified), plus the standard contractual clauses in the Google Workspace data processing terms.
  • OpenAI: our data processing agreement is with OpenAI Ireland Ltd.; transfers to the United States are covered by the European Commission’s standard contractual clauses.

If you want a copy, email info@retrieva.eu.

7. How long do we keep your data?

DataRetention period
Account and everything you made and uploadedAs long as you keep your account. If you do nothing in the app for 24 months, we automatically delete your account and everything in it. About 30 days beforehand we email you about this, if we have a working email address. An active paid subscription is never cleaned up.
After you delete your accountImmediately erased from the database and from disk, including photos and snapshots of your account. If an administrator deletes your account, for example at your request, we keep a snapshot for another 30 days so it can be restored; after that it is gone. We make one backup of the whole system per day, keep it encrypted (AES-256) and for no longer than 10 days; after that your data is no longer in them either.
Payments and invoices7 years, because tax law requires us to. Once your account is deleted we detach payments from you; invoices are kept as the law requires.
Agreement and consentAs long as your account exists, and afterwards as long as needed to prove that you agreed.
Login sessionsUntil you close your browser (at most 24 hours), or 7 days if you tick ‘Remember me’; shorter if you log out.
Site visitor countsNo personal data; we discard the daily counts after about 13 months.
Origin data (including campaign, ad and how you signed up)As long as your account exists; deleted together with your account.
“Isn’t this right?” reportsAs long as your account exists; deleted together with your account.
Reason for cancellingAs long as your account exists; deleted together with your account.
Reason for deleting your accountAt most 24 months, without any link to you (anonymous).
Which emails we sent you (to which address, the subject and whether it arrived; not the content)6 months, so we can check whether an email reached you; deleted sooner if you delete your account.
How many AI actions you used per day and what they costAs long as your account exists, so we can safeguard the limits and follow the costs; deleted together with your account.
Code of the internet connection when an account is created7 days.
Hits of the content check (only the category and the time, never the content)90 days. An automatic block stays until we have reviewed it.
When you were active in the app, per hour90 days. We only use this to see which hours are busy; deleted sooner if you delete your account.
Payment attempts that were not completed (cancelled, failed or expired)30 days. A successful payment is kept for 7 years (see above).
Messages to us (emails and what you write to us)As long as needed to handle your question, and then at most 2 years, so we know the background if you ask a follow-up question.
Log of what administrators do (for example a withdrawal, a refund or a change to your account, with your name and account number)12 months, so we can check who did what and when.

8. How do we protect your data?

  • All connections use https.
  • Passwords are never stored readably: only as a hash (scrypt; bcrypt for older accounts).
  • An API key you enter yourself is stored encrypted (AES-256-GCM).
  • Uploaded photos and files can only be retrieved by you, logged in; sharing a link does not give access.
  • Login cookies cannot be read by scripts (httpOnly) and are not sent to other sites.
  • The site loads no scripts, fonts or images from other parties, so no data leaks to third parties just by opening a page.
  • Limits on login attempts, daily backups that are stored encrypted (AES-256), and administrator access only where necessary.

Despite all care, a data breach can happen. If it does and there is a risk to you, we will report it within 72 hours to the Dutch Data Protection Authority and, if the risk is high, to you as well.

9. Your rights

Under the GDPR you have the following rights:

  • Access: knowing which data we hold about you.
  • Rectification: having incorrect data corrected. You change your name, email address and password yourself in the app.
  • Erasure: having your account and all your data erased. You do that yourself in the app under Settings; it happens immediately.
  • Restriction and objection: objecting to processing based on our legitimate interest, for example the origin data.
  • Portability: taking all your data with you. In the app you can download a zip under Settings with all your data (JSON, plus CSV, ICS and Markdown) and your uploaded photos and files, and your question bank and flashcards separately as well. Only your password and the secret keys in your links are left out.
  • Withdrawing consent: for example for the weekly summary. You switch this off in Settings, or via the unsubscribe link at the bottom of every email. Reminders are switched off the same way. Earlier processing remains valid.

If something does not work in the app, email info@retrieva.eu. We respond within 2 working days, and handle your request within one month at the latest, as the GDPR requires. To be sure it is you, we may ask for confirmation.

Do you disagree with how we handle your data? Please contact us first. You also always have the right to lodge a complaint with the Dutch Data Protection Authority (Autoriteit Persoonsgegevens, autoriteitpersoonsgegevens.nl) or your local supervisory authority.

10. Cookies and local storage

We only use cookies and storage that are necessary for the app to work, and no advertising or tracking cookies. See the cookie statement.

11. Changes

If something important changes in how we handle data, we will update this statement, raise the version date at the top and tell you in the app or by email. This statement is version 2026-11-16.1. The Dutch version is authoritative.

DPIA