Opening your learning space
Starting the app …
Opening your learning space
Starting the app …
pursuant to Article 13 GDPR
Complete English translation of the published German Privacy Policy, version 3.4, dated 16 October 2026. References to German and European law are retained from the published original.
AM Creative Tech UG (haftungsbeschränkt)
Herweghstr. 8, 01157 Dresden, Germany
Telephone: +49 151 15434471
Email: amos@am-creativetech.com
Managing Director: Amos Matzke
When you visit the website, we collect the following data, which your browser transmits automatically:
The data is stored in our system's log files. We do not combine this log-file data with your user account. This is separate from application error reports, which contain similar technical details and, for signed-in users, are linked to the account identifier; these are described in section 2a.
If a technical error occurs in the application, your browser creates an error report and sends it to us. If a request to an AI service fails, our server also creates an error report containing your account identifier, the function requested and the error message; it contains no device or browser details. A browser report contains the error message and the sequence of calls within the program (which may, in individual cases, include extracts from an API response), the name of the screen and function in which the error occurred, the path of the requested page and the previously visited page without the query string or URL fragments (so one-time verification codes do not enter the report), details of the browser, operating system, language, screen and window size, pixel density, colour scheme, connection type, memory usage and length of your visit, and, if a server request fails, the address, status code and duration of that request and the AI model's name. If you are signed in, it will generally also contain your account identifier, which is also part of the report's identifier. If the usual reporting route fails, for example because the error affects sign-in itself, we store the report without an account identifier. If you are not signed in, it never contains one. We store your email address in a report only if you write feedback after an error (up to 400 characters), so that we can ask follow-up questions. If the same error occurs again on the same screen, we increase the count in the existing report instead of creating a new one.
The purpose is to identify and resolve errors, ask follow-up questions on feedback, and allow the administration to view the number of new errors per day and an account's most recent reports. The legal basis is Article 6(1)(f) GDPR; our legitimate interest is the stable operation of the application. Error reports are stored in Google Cloud Firestore in the europe-west3 region (section 3) and are accessible to us; you receive your own reports as part of the access request described in section 8. They are deleted automatically 90 days after the error last occurred, and immediately when your account is deleted. You may object to this processing under Article 21 GDPR; please write to the address given in section 1.
When you register, we collect the following required information:
Processing is based on Article 6(1)(b) GDPR (performance of a contract). Your registration data is processed and stored by Firebase Authentication (Google LLC) in the United States. Learning-content data (flashcards, topics and learning progress) is stored in Google Cloud Firestore in the region europe-west3 (Frankfurt, Germany) . Google LLC is certified under the EU-US Data Privacy Framework.
When you create an account, we record in your user account how you came to us. We store only:
/karteikarten-mathe ) utm_* , gclid ), if present The purpose is to evaluate which of our content and advertisements actually lead to registrations. Without this information, we would not be able to assess which pages we should develop further. It is not used for profiling or targeted advertising. We do not disclose the information stored in your account to third parties. Independently of this, the shortened entry path and the classification of the referring page into a channel are included in the cookieless audience measurement described in section 6b, but campaign parameters are not.
Without your consent, we do not access your device: In that case, this information is held only in your browser's working memory and disappears when you close the page. No cookies are set for this purpose, and nothing is stored in or read from your device's local storage. Consent under section 25 TDDDG is not required for this route. The information becomes part of your account data only when you create your account. To retain it during sign-in via Google or Apple, it may temporarily appear as a visible parameter in the sign-in page's address bar.
With your consent to web analytics, we also store it on your device: If you have consented to the web analytics described in section 6, we store the same information together with the GA4 client ID mentioned below under the key quanta_first_touch_v1 in your browser's local storage and read it again on subsequent visits. This is the only way to retain a first contact from before your current visit. This access to your device is based solely on your consent. If you withdraw it, we stop reading and updating the entry and delete it from your device when you withdraw consent.
Legal basis: Article 6(1)(f) GDPR (legitimate interest in measuring the reach and effectiveness of our own services) and, when an account is created, also Article 6(1)(b) GDPR. Your consent is the basis for storing and reading information on your device as described in the preceding paragraph (section 25(1) TDDDG, Article 6(1)(a) GDPR). The information forms part of your user account and is removed when the account is deleted. The entry in your browser's local storage is separate: we delete it when you withdraw analytics consent, and you can remove it yourself at any time through your browser's website-data settings. You may object to processing under Article 21 GDPR; please contact the address given in section 1.
This is separate from the Google Analytics 4 web analytics described in section 6. Only if you have consented to those analytics is the GA4 client ID also stored in your account, so that a subsequent purchase can be attributed to the original source. Without your consent, this part does not take place at all.
Contact through the first-year student promotion. If you reach us through our promotional page /ersti-angebot , for example via the QR code on a flyer, we also record this once in your account, including if you already have an account. We store only:
utm_source and utm_campaign , if they are present in the address accessed Only the first contact counts; subsequent visits do not overwrite it. The purpose is to evaluate how many accounts reach the promotion through the flyers and promotional page, and to preselect the free month with the code Ersti26 in the offer window. No cookies are set for this either, and nothing is stored in or read from your device's local storage: the information is derived from the address accessed and, where that does not identify the source, from the referring website if your browser transmits it. We retain only its classification from the list above, never the referring website's address; the information is stored directly in your account. If you enter the prize draw, we copy the route, source and campaign parameters, without the timestamp, into your entry record (section 5b). As above, the legal basis is Article 6(1)(f) GDPR. Both purposes end with the promotion; we therefore delete this information at the end of 31 March 2027 together with the entry records, or at the latest when your account is deleted. You may object under Article 21 GDPR by contacting the address given in section 1.
Personal access for ChatGPT and other suitable AI assistants via MCP is currently being prepared and has not yet been released publicly. The following description concerns this new route, which can be used only after a separate connection to your account; it does not automatically change the data-sharing settings of existing accounts. Managed school and institution accounts are not enabled for this personal access.
You sign in to Quanta and expressly authorise the requested permissions to read your own learning materials, save confirmed drafts or record your actual learning answers. The assistant does not receive your Quanta password. Quanta stores the link to your account identifier, the authorised assistant, its permissions, validity and revocation details, and hashed technical access credentials in the existing server-protected account and database areas. Access credentials are not exported as part of your learning content.
Depending on the permission granted and function requested, the selected AI client receives your account identifier and, where applicable, your display name, as well as your own topics, cards, confirmed learning progress, selected source passages and their document or page references. Quanta does not automatically transfer a complete uploaded file to the client through this route. The client may in turn process the retrieved content to generate its model response; the information provided by your chosen provider applies to its subsequent processing, recipients, processing locations and storage periods. Connecting an account does not entitle you to any particular settings from that provider.
Cards drafted by an external client are added to your personal Quanta collection only after you confirm an instruction to save them. They remain labelled as external drafts that Quanta has not checked. Actual card ratings and multiple-choice answers are stored after your confirmed learning action, together with the necessary result record and protection against duplicate processing. This does not automatically give Quanta your entire conversation with the assistant.
New generation by Quanta may begin only after your confirmed instruction, the required account permissions and technical authorisation are in place. For this purpose, the topic description associated with the instruction and selected, already processed source passages that you are entitled to use are transmitted to Mistral AI through the existing text-processing route, as described in section 4. Quanta maintains the instruction, source, status, saved result and technical cost reservation to carry out the request and prevent duplicate work. This does not enable direct file import or OCR from the assistant.
You can revoke individual connections in your Quanta profile. Expiry, revocation, account suspension or account deletion ends further access under the applicable account agreement. Revocation alone does not delete your own learning materials that have already been saved. Account deletion includes the associated connection and request data held by Quanta under the existing deletion policy; aggregate operational counters across accounts do not contain an account identifier or learning text.
Revoking a connection in Quanta does not retroactively delete content from an external client you have already used or from its conversation history. Use that provider's deletion and privacy functions for this purpose. A source link or saved result is not a promise about that provider's privacy practices, automatic publication or a recommendation by Quanta.
When you open a personal offer, we keep the necessary account and payment association, quoted terms and deadlines, and display, decline, payment-command and actual acceptance states on the server. This lets us resume promised offers, prevent duplicate charges and avoid issuing limited benefits repeatedly to the same verified identity. Accepting an offer does not subscribe you to newsletters or advertising.
Processing your requested offer and contract relies on Article 6(1)(b) GDPR. The necessary, limited prevention of repeated redemption relies on Article 6(1)(f) GDPR; our interest is fair allocation of limited benefits and prevention of abusive repeat claims. Learning content is not analysed for this purpose.
On account deletion, we remove account-linked offer history after safely reconciling open payments. Only an actually accepted benefit may leave a necessary minimal receipt: a secret-keyed check value of the verified identity (HMAC), previously granted benefit phases and the associated policy and expiry records. This value is pseudonymous, not anonymous. The unlinked receipt contains no email address, account ID, Stripe ID, learning content or marketing consent, and cannot grant account access.
The minimal receipt is kept only while necessary, for no longer than three years after the end of the year of the last accepted benefit phase. Rechecking does not extend its fixed expiry. Unneeded receipts and histories without continuing retention approval are deleted. Statutory invoice records at Stripe remain subject to the separately described rules.
Account-linked history is included in your data access/export. You may request access, correction, erasure, restriction and objection under Article 21 GDPR through our privacy contact, including after account deletion; we verify your identity and review the request. An objection to advertising is separate. The minimal receipt is not used for newsletters or other advertising.
When you create or join a private study group, write a message or share study material, we process its name, account identifiers, membership and invitations, your messages, timestamps and your own read state. Members see contributions intended for the group and the associated profile names. Until joining, invitees see only their invitation and information about the inviting group owner, not earlier group contributions or the full member list.
Study material is referenced, not copied into the conversation. Viewing and importing still depend on the valid share, its expiry and your membership. Revoked or expired shares no longer reveal source content in the timeline. Copies previously imported into a recipient's own learning area remain independent of the original share.
New messages and conversation events become unavailable for display and export 30 days after creation. Firestore then deletes them asynchronously. Reading or reopening does not extend this period. The group name, membership, open invitations and content-free read state remain while the group exists, until removed or deleted. Existing content-share expiry periods do not change.
Deleting a group actively removes its messages, content shares and personal conversation views. Account deletion also removes your own contributions and invitation events in other groups, including groups you have already left. A minimal deleted-group marker prevents accidental recreation by repeated requests until the owner's account is deleted; it contains no group or message text and no other members' roster.
We process these data to provide the private group function you request under Article 6(1)(b) GDPR, within the existing Firebase infrastructure. No additional external messaging provider is used, and group messages are not analysed by AI. Your retained contributions and read states form part of your data access/export. A short email may notify you of a new group or friend invitation unless invitation emails are disabled in your notification settings. Its link opens the signed-in acceptance view and never accepts automatically. The general information on recipients, security, backups and your rights continues to apply.
Quanta uses AI services to process learning content. The definitive overview of the provider, purpose and basis for transfer is given in section 16. The following data is transmitted to the AI service:
Only the learning content necessary for the relevant function is transmitted to the AI service. No direct account data (name, email address, IP address or user identifier) is passed to the AI service. Processing is based on Article 6(1)(b) GDPR (performance of a contract) and Article 6(1)(a) GDPR (consent for optional AI use).
The AI service provider is Mistral AI SAS, 15 rue des Halles, 75001 Paris, France (Paris Commercial Register no. 952 418 325). Processing takes place through this provider's API. For text processing (card generation, multiple-choice tasks, learning tutor, exam simulation and formula conversion), we use the provider's regional EU endpoint; according to its documentation (accessed on 2 October 2026), inference processing takes place in the EU or EFTA; control and administrative data may be processed outside this region. For automatically reading uploaded documents and photos and transcribing voice recordings, we currently use the standard endpoint, for which the provider does not contractually guarantee a specific processing location.
According to the provider's public list of service providers (as of 18 August 2026), its processing infrastructure is predominantly located in the European Union and the European Economic Area, including France, the Netherlands, Belgium, Sweden and Norway. Some of its service providers, such as a global network and delivery service, also process data outside the EU. These transfers are based on the provider's data processing agreement, including the European Commission's standard contractual clauses under Implementing Decision (EU) 2021/914. You can obtain a copy of these safeguards by writing to amos@am-creativetech.com .
In its terms of use, the provider undertakes not to use content transmitted through its API to train its models. We have also opted out of the use of our data to improve the provider's services in its administration settings. We do not submit feedback to the provider and do not use models labelled 'Labs' or 'Preview', because the provider's undertaking expressly does not apply to these two cases.
For stateless AI requests covered by our activated provider setting, the provider stores inputs and outputs no longer than necessary for processing. This includes chat, embeddings, OCR and transcription. Uploaded files and batch processing are excluded. We delete transferred file copies once processing is completed or cancelled; deletion is also attempted if an error occurs, and a regular cleanup process subsequently deletes any copies left behind. Batch processing is subject to the separately described maximum period of eight days. Our own storage of your learning content is independent of this provider setting. Voice recordings are transmitted directly to the transcription API; no file is stored there.
We also use Google services for the podcast function's text-to-speech output and its Hebrew speech recognition; details are provided in section 4e.
In the 'Documents' area, you can upload your own learning documents (PDF files or images) to store them permanently in your account and generate an AI-assisted summary. The following is processed:
The purposes are the lasting, readily accessible storage of your learning documents, automated summaries to support learning and, optionally, the creation of flashcards or quiz questions from a document. The legal basis is Article 6(1)(b) GDPR (performance of a contract); section 4 applies accordingly to optional AI use.
The metadata and summary are stored in the same Firestore database (europe-west3 region, Frankfurt am Main) as the other learning content; the file itself is stored in Firebase Storage in the europe-west1 region (Belgium) within the same Google Cloud project. To create the summary, the document's content is transmitted to the AI service named in section 4; the information in section 4 applies accordingly (no automatically included core account data, no training with your content). If the document does not contain an extractable text layer, a copy of the file is also transferred to the provider's file API, read automatically there and deleted by us immediately after processing. Your stored document is unaffected.
Storage periods and your rights: Documents removed individually are initially moved to the trash for 30 days and can be restored during that period. They continue to count towards storage during this time. Explicitly confirmed permanent deletion removes the original file, summary and derived text artefacts from active storage; private source references in existing cards and tests remain as provenance information. Subsequent automatic cleanup of original documents in free accounts that have been unused for at least 24 months is not currently activated. Before activation, the rule and its introduction will be communicated. Deletion will take place no earlier than 90 days after a personal advance notice has been successfully delivered; reminders will follow 30 and seven days before deletion. Active paid subscriptions, trials and school accounts are excluded. New use stops a scheduled cleanup. Missing usage history will be treated conservatively from the time of introduction; after paid access ends, a grace period of at least 90 days also applies. This does not delete learning content, the account or billing records. Account deletion removes the active collection through the existing resumable deletion procedure; the 30-day trash period does not delay it. Technical soft-delete and backup periods are unaffected and are described in this Privacy Policy. Soft delete keeps deleted files technically recoverable for 7 days. Native database backups are retained for 7 days; daily database exports are automatically deleted once they are 30 days old and remain recoverable through soft delete for another 7 days. These copies cannot be accessed through the app. Lifecycle deletion is asynchronous and does not guarantee an exact time of physical deletion. The data export (Article 20 GDPR) contains the metadata of all documents and the full text of their summaries; original files can be accessed and downloaded directly in the app at any time.
In the 'Get started' section of the homepage, you can enter a topic or place a document (PDF or image) before you have an account. Both remain exclusively in your browser: the topic in session storage (sessionStorage), and the file itself in the local browser database (IndexedDB) of the same tab. No data is transmitted to us or to third parties at this stage. The file is not uploaded to a server or sent to AI, and does not appear in any server log, URL or analytics event. The file name is not reported anywhere either.
The sole purpose is the function you requested: continuing with exactly this input after registration. Storage on your device is based on section 25(2) no. 2 TDDDG (strictly necessary to provide the service expressly requested); at this stage, we do not process personal data within the meaning of the GDPR because the content does not reach us.
Storage period and deletion: A file placed there is automatically deleted from the local browser database no later than 90 minutes after being placed there, even if you abandon registration. You can remove it yourself at any earlier time using the remove icon in the 'Get started' section or 'Do not use' on the registration page. It is uploaded through the regular document process only when you create an account and onboarding assigns the file to your first topic; section 4a applies from then on. If you clear browser storage or use private browsing, the file may not be stored; the homepage then informs you and offers an upload after registration.
To find suitable source passages for AI-assisted learning functions, we divide stored personal documents into text passages during processing. Mistral AI SAS uses these to compute a vector representation (embedding), a sequence of numbers for comparing similarity of meaning, through its regional EU endpoint. Search queries are also converted into vector representations for this purpose. The information about the AI service in section 4 applies, including the activated 'no data retention' option for the stateless embedding API and the exclusion of training use. To search sources in other languages, search queries may be translated by the same AI service; verbatim evidence remains in its source language.
In your personal account, we store the text passages, their page and position references, technical checksums and version identifiers, and their associated vector representations in Firestore (Frankfurt am Main). This also applies to teachers' own documents that have additionally been shared with a school: the passages and vectors remain in the personal account and do not thereby become accessible to the school or other accounts. Documents in the school area and accounts with the 'learner' role are excluded from this storage.
The legal basis is Article 6(1)(b) GDPR (provision of the requested learning functions); for optional AI use, the information on consent in section 4 also applies. The passages and vectors are deleted with their associated document and when the account is deleted. They are included with the document in the data export under section 8.
If you dispute a source reference, we store the reason selected, your optional note (up to 500 characters), the disputed reference and a technical run identifier for attribution. This information is used to review and improve source selection. The legal basis is Article 6(1)(f) GDPR; our legitimate interest is troubleshooting and quality assurance. The reports are stored in the personal account in Firestore (Frankfurt am Main), are accessible to authorised administrators, are included in the data export and are deleted when the account is deleted.
AI functions for podcasts are currently enabled only for accounts for which we have expressly activated them: Quanta administrator accounts and accounts of licensed institutions. Other accounts can also make their own private recordings without AI processing, within their storage and recording limits. Depending on the functions enabled, you can have a script generated from selected personal documents and learning content, have it voiced by AI voices (text-to-speech), make a recording yourself or add your own audio file, and have a recording converted into text (speech-to-text). The following is processed:
The purpose is to provide the function you requested. The legal basis is Article 6(1)(b) GDPR (performance of a contract). Recording begins only when you start it and your browser permits access to the microphone; a transcript of your recording is created only at your request. We do not analyse voices to identify people. If you record other people, you need their agreement (section 8b of the Terms of Use). For institutional accounts, we process this data on behalf of the institution (section 14).
Storage locations and recipients: We store recordings, audio tracks, scripts and transcripts in your account in Firebase Storage in the europe-west1 region (Belgium), and the associated information in Firestore (Frankfurt am Main). A dedicated processing service in our Google Cloud project in the europe-west4 region (the Netherlands) converts and checks audio files. The Mistral AI service named in section 4 generates the script from your selected sources through its regional EU endpoint. In addition:
When AI voices are used, the approved script remains the text associated with the recording. The audio track is not transcribed again or subsequently checked against the script. If you record an approved script yourself, it remains the text you supplied; the spoken content is not automatically checked. The information on the transfer basis in section 16 applies to Google's services. Learners in institutions can only make their own recordings; scripts, AI voices and transcripts are not available to them, and we do not transmit their recordings to any of the speech or AI services named above.
Storage period and deletion: An episode's recordings, audio tracks, scripts and transcripts remain stored until you delete the episode or your account is deleted. Deleted episodes initially remain in the trash for 30 days, during which you can restore or permanently delete them; a regular cleanup process then deletes them and their files. If an episode is voiced again, we retain the most recently replaced audio track for 30 days. We delete further replaced versions, a recording's uploaded original file and intermediate processing files, including the copy for Hebrew speech recognition, 24 hours after processing finishes. A version shared within an institution is retained until that sharing ends. Deletion processes run asynchronously and do not promise an exact deletion time. When an account is deleted, we remove all of its podcast data and cancel and delete ongoing recognition operations at Google. The 7-day soft-delete period described in section 4a applies to deleted files in Firebase Storage. Scripts, transcripts and information about your audio files are included in the data export under section 8; you can download the audio files themselves in the podcast view.
We use the payment service provider Stripe, Inc. (510 Townsend Street, San Francisco, CA 94103, USA) to process payments for Quanta Essential.
When you take out a paid subscription, the following data is transmitted to Stripe:
Stripe processes this data as an independent controller in accordance with the Stripe Privacy Policy . The legal basis is Article 6(1)(b) GDPR (performance of a contract). Stripe is certified under the EU-US Data Privacy Framework.
A separate item is an attribute that we ourselves record against your subscription at Stripe: your request for performance to begin immediately. Before checkout begins, you confirm that we should begin providing the service immediately (section 6 of the Terms and Conditions). We store the time of this confirmation and the version of the Terms and Conditions then in force as attributes of the checkout and subscription at Stripe, and in your user account together with the checkout identifier. The legal basis is Article 6(1)(b) GDPR, because your declaration forms part of entering into the contract, and additionally Article 6(1)(f) GDPR, based on our legitimate interest in being able to prove when performance began in the event of a dispute; you may object to this additional processing under Article 21 GDPR. We delete the information in your user account when the account is deleted. The attribute held at Stripe remains for as long as Stripe retains the subscription record; we stop evaluating it after the end of the third year following the year in which the contract was concluded (sections 195 and 199 BGB).
Promotional codes. If you start checkout using a code from the first-year student promotion ( Ersti26 or Semester26 ), clicking accepts the promotional terms. We store the version then in force and the time of acceptance in the same way as the declaration requesting immediate performance: as attributes of the checkout and subscription at Stripe, and in your user account together with the checkout identifier, under the same legal bases and retention periods. This does not include information about the prize draw (section 5b).
Contract confirmation by email. After you take out a paid plan, we send a contract confirmation to the address in your user account, including the plan, term, price, trial period, information on the right of withdrawal, model withdrawal form and your declaration requesting immediate performance (section 312f(2) BGB, Article 6(1)(c) GDPR). We record in your user account which checkouts have already been confirmed, so that confirmations are not sent twice. Delivery: section 6g.
You can cancel a paid contract without signing in using the 'Cancel contracts here' button in every page's footer (section 312k BGB), and withdraw from it within the withdrawal period using the 'Withdraw from contract' button (section 356a BGB). We process the information you enter in the respective form: for cancellation, the type of cancellation, the reason where applicable, and the requested termination date; for both, your name, your account's email address, and the date and time your declaration is received. We use this email address to associate the declaration with your account and contract at our payment service provider, carry out the cancellation or withdrawal there, and promptly confirm receipt and the outcome by email to that address. If a declaration cannot be matched automatically, we review it manually. Your declaration is considered received when the form is submitted. The legal basis is Article 6(1)(b) and (c) GDPR in conjunction with sections 312k(4) and 356a(4) BGB. We retain the declaration and its time of receipt for three years, calculated from the end of the year, to be able to prove receipt (sections 195 and 199 BGB); we then delete them.
Until 30 November 2026 , we will draw one winner from all valid entries for Apple AirPods 4 with active noise cancellation (ANC) . Details are set out in the Terms of Entry . We process your data to record your entry, check eligibility and duplicate entries, conduct and document the draw, notify the selected person and send the prize.
You can enter in two ways, and we collect data differently depending on the route:
We process:
The draw on 3 December 2026 uses a random procedure; there is no assessment of you as a person. We record the time, the number of entries, a verification hash of the entry list, the randomly drawn order of entry identifiers (which contains your account identifier for the free-month route) and, for each person whose turn is reached, the times of notification, acceptance, forfeiture, exclusion with its reason, and shipment, as well as their email address while it is their turn.
The legal basis is Article 6(1)(b) GDPR: processing is necessary to conduct the prize draw under its Terms of Entry. We retain the draw record even after your account is deleted, based on our legitimate interest in being able to demonstrate a fair draw and the award of the prize (Article 6(1)(f), Article 17(3)(e) GDPR).
Recipients. We do not disclose prize-draw data to Stripe. Stripe's checkout record states only that this is the promotion's free month, when it ends, where in the application you began checkout, and which version of the promotional terms you accepted and when (section 5); whether you enter the prize draw, your age information and your residence are not sent to Stripe. Entry data is held in our Google Firebase database (section 3). Google Ireland Limited operates our Google Workspace email inbox as our processor; it holds emails for entry without purchase and the selected person's reply containing their name and shipping address (section 16). We disclose the selected person's name and shipping address only to the parcel or postal service delivering the prize.
Storage period. We delete entry records at the end of 31 March 2027 , and also delete the emails for entry without purchase. If you delete your account earlier, we delete your entry with the account and your participation ends; after the draw, only your entry identifier remains in its record. We delete the email containing your name and shipping address after delivery; at the same time, we remove the email address from the draw record and shorten the drawn order to those people whose turn was reached. If a person's turn ends earlier because the prize was forfeited or they were excluded, we remove their email address at that point. We delete the draw record itself 6 months after the prize is delivered, or, if no delivery takes place, 6 months after the final step in administering the draw.
Your rights under section 8 apply here too; your entry is included in your account's data export. You can end your participation at any time before the draw by writing to the address given in section 1 or deleting your account. This does not prevent you from using the free-month promotion: the free month continues without participation in the prize draw.
Quanta uses Google Analytics 4 (GA4), provided by Google Ireland Limited, Gordon House, Barrow Street, Dublin 4, Ireland. GA4 is activated only after your express consent through the cookie banner ( Consent Mode v2 ). Without your consent, the GA4 script is not loaded at all; no connection to Google is established, and no cookies are set or measurement pings sent. If you later withdraw consent, we also report the state denied to Google and no longer load the script on your next page visit.
If consent has been given, the following data is processed:
Where GA4 measures activity. GA4 is not loaded in the school or administration areas of our application, even with your consent; no connection to Google is therefore established there. The same applies to pages whose address may contain a one-time action link, namely verification and password links and the matching of a cancellation request. In contrast, the page where you set up your personal account immediately after registration and the page asking you to verify your email address are not excluded; GA4 measures activity there if consent has been given, because their addresses contain no code. If you move from a public area into the school or administration area without a new page load, GA4 is disabled for the rest of your visit. In contrast, GA4 does measure activity in the signed-in learning area, unlike cookieless audience measurement (section 6b) and unlike the Meta Pixel, which does not report page views there (section 6d). It transmits the address of the page visited, not its content; the address may contain technical identifiers for the function opened, such as the name of a formula accessed. Your cards, notes and inputs are not transmitted.
Legal basis: Article 6(1)(a) GDPR (consent). You can withdraw consent at any time through the cookie banner. Google Ireland Limited is certified under the EU-US Data Privacy Framework. Further information: Google Privacy Policy .
On our public pages (homepage, information and funnel pages), we also use Umami , a cookieless, data-minimising audience-measurement service provided by Umami Software, Inc. (USA). Umami sets no cookies and stores no persistent recognition identifier on your device. According to the provider, Umami counts visits anonymously using a hash that changes daily , intended to prevent recognition across days; this hash is generated by the provider, not on your device. No personal data is stored and no profiles are created.
The signed-in area and the exceptions. Umami does not count page views in the signed-in area (learning app, school area or administration); it does count two pages in the sign-in flow—setting up your personal account after registration and the request to verify your email address—without requiring consent. Two groups of events are exceptions. The first consists of events describing a completion that must also be counted after a redirect to the signed-in area: completed registration, the start of checkout and completed checkout. The second consists of three exam-simulation events (session start, cancellation and completion), which let us identify where sessions are abandoned. Both groups are entirely excluded in the school and administration areas; nothing is counted there. Only the event name and a few fixed details are reported: the selected sign-in method, plan name and amount, whether it concerns a trial, a payment not yet received or a completed paid purchase, the general part of the application where the purchase occurred (such as the generator or paywall), and the entry page and channel for your visit. Instead of the address of the page you are currently on, a fixed placeholder ( /app ) is included; the checkout number is not included.
Only anonymous usage data is processed:
Umami does not store anything on your device. The only access to local storage is a read operation by the provider's script before each transmission: it checks the key umami.disabled , which can be used to disable measurement permanently on a device. Quanta never writes this key, so it will generally not exist for our users; what is read is therefore neither device nor user information, but only whether you have disabled measurement yourself. Because no other information is stored on or read from your device (section 25(2) TDDDG) and no personal data is processed, the service is used without consent . The legal basis is our legitimate interest in data-minimising, cookieless audience measurement (Article 6(1)(f) GDPR). Further information: Umami Privacy Policy .
To protect against misuse, automated attacks and excessive API usage, Quanta uses the Upstash Redis service provided by Upstash, Inc. (340 S Lemon Ave #1743, Walnut, CA 91789, USA). Your IP address is temporarily processed in Redis storage (between 60 seconds and one hour, depending on the protective purpose) to count the frequency of requests. On the cancellation and withdrawal pages (section 5a), we also count an irreversible checksum of the email address entered so that nobody can flood the same address with confirmations; the address itself is not sent to Upstash. The IP address is not stored permanently or linked to your user account.
Legal basis: Article 6(1)(f) GDPR (the provider's legitimate interests in the platform's security and stability). Upstash processes data in data centres within the EU (Frankfurt). Further information: Upstash Privacy Policy .
Firebase Authentication remains responsible for identity and for generating and verifying one-time verification and password links. The resulting account, security and consent emails are delivered through the transactional email service Resend (Resend, Inc., 2261 Market Street #5039, San Francisco, CA 94114, USA):
For secure operation, we store internally only redacted delivery statuses and an irreversible, secret-based verification signature for bounce, complaint or provider-block notifications. This prevents further automatic delivery attempts; the actual email address, action links and email contents are not stored in the delivery queue or operational logs. We do not use open or click tracking.
The legal basis for account, security and consent messages is Article 6(1)(b) GDPR and, where necessary, Article 6(1)(f) GDPR (secure account administration and delivery). Adult B2C users can separately manage an optional setting for future learning and product information in their account settings. This consent is not part of consent to the Terms and Conditions or Privacy Policy; its version, language, source and time are documented, and it can be withdrawn at any time. No optional product or lifecycle emails are currently sent under this setting. Further information: Resend Privacy Policy .
To protect against automated misuse, bot access and fraudulent use, Quanta uses Google reCAPTCHA Enterprise together with Firebase App Check (Google LLC) in the signed-in app area. Browser interaction and device signals (such as usage and movement patterns and device and browser properties), together with your IP address, are transmitted to Google and processed there to produce a risk assessment. This assessment runs invisibly in the background; no CAPTCHA puzzle is displayed and no separate action is required from you.
The legal basis is Article 6(1)(f) GDPR (legitimate interest in preventing misuse and fraud and in ensuring the platform's security and stability). The recipient is Google LLC (1600 Amphitheatre Parkway, Mountain View, CA 94043, USA); Google LLC is certified under the EU-US Data Privacy Framework. Further information about Google's data processing can be found in the Google Privacy Policy .
If you enable the category Marketing in our consent banner, we integrate the Meta Pixel provided by Meta Platforms Ireland Limited (Merrion Road, Dublin 4, D04 X2K5, Ireland). Its sole purpose is to measure our advertising campaigns: we want to identify which advertisement led to a registration or subscription. The category is disabled by default . Without your consent, the Meta script is not loaded ; no connection whatsoever to Meta is then established, and no data is transmitted or information stored on or read from your device.
There are two routes , and both depend on the same consent. The first is the pixel in your browser, discussed in the following paragraphs. The second is a report from our server for precisely one event: a completed paid purchase. This second route is in operation ; what it transmits is described below under 'Reporting a completed paid purchase'. No transmission takes place through that route without your consent either.
Where the pixel does not run. The Meta script is not loaded in the school or administration areas of our application, even with your consent; no connection to Meta is therefore established there either. If you move there from a public area without a new page load, the script has already been loaded in your browser; it is then disabled for the rest of your visit and no longer reports anything from these areas to Meta. In the signed-in learning area, it loads but does not report page views to Meta, only the registration, offer, checkout, trial and completed paid-purchase events listed below. Meta does not learn which learning content you open there.
If consent has been given, the following data is transmitted to Meta:
Reporting a completed paid purchase (Conversions API). When a paid purchase is completed, we also report that completion to Meta from our server using the Conversions API. The reason is reliability: a report from your browser is lost if you close the completion page beforehand or a browser extension blocks it. Both reports carry the same event identifier so that Meta still counts the purchase only once.
This report is sent only if marketing consent has been given . So that our server can check this afterwards, we additionally record your response to the Marketing category in your user account together with the time of the decision. Withdrawing consent overwrites this record immediately, and no report is sent without a record of granted consent. The record is deleted with your account.
Only the following is transmitted:
What is not transmitted through this route: no email address, name, telephone number or postal address, including in encrypted (hashed) form. Your IP address and your user identifier with us are not transmitted either. We could improve Meta's matching rate with these details, but deliberately do not do so because they would make it possible to attribute the completed purchase to a person known by name.
We do not want automatic advanced matching , in which the pixel reads inputs from form fields. We have disabled it to the extent within our control: the pixel's automatic data collection is disabled in our page source before the pixel starts. Meta provides no further switches for this in the page source. Independently of that, we never supply customer data to the pixel: we send no email addresses, names, telephone numbers or postal addresses to Meta, including in encrypted (hashed) form.
What we cannot guarantee: when it is accessed, the pixel loads its settings from Meta's servers. What those settings permit is outside our source code. We therefore describe the measure we have taken here without promising a result we cannot verify ourselves.
The legal basis for storing and reading information on your device is section 25(1) TDDDG; the legal basis for subsequent processing and transfer of your personal data is Article 6(1)(a) GDPR (consent).
Transfer to the United States. Meta also processes the data in the United States of America through Meta Platforms, Inc. Meta Platforms, Inc. is certified under the EU-US Data Privacy Framework; the European Commission's standard contractual clauses also apply. Nevertheless, a level of protection fully equivalent to European standards cannot be guaranteed. In particular, access to the data by US authorities cannot be ruled out. By giving consent, you also consent to this transfer (Article 49(1)(a) GDPR).
Joint controllership. We and Meta Platforms Ireland Limited are joint controllers within the meaning of Article 26 GDPR for collecting data on this website and transmitting it to Meta. This joint responsibility is limited to that precise step. Meta alone is responsible for everything it subsequently does with the data. Meta makes the relevant agreement available here: Controller Addendum . You may exercise your rights as a data subject against either us or Meta.
Withdrawing consent. You can withdraw your consent at any time with effect for the future. To do so, open the cookie settings and disable the Marketing category or select 'Necessary only'. The settings are available in our website's footer and, when signed in, also in your profile under 'Data & Privacy'. The pixel stops transmitting immediately and is no longer loaded on your next page visit. If you are signed in, the withdrawal is also recorded in your user account; this also prevents the server-side report of a completed paid purchase, including for a checkout still in progress at that time. The lawfulness of processing before withdrawal is unaffected; we cannot retrieve data already transmitted to Meta. The cookies previously set, _fbp and _fbc , remain stored on your device after withdrawal; withdrawing consent does not delete them. They are no longer evaluated without the pixel being loaded, and you can remove them at any time through your browser settings. Further information: Meta Privacy Policy .
In operation since 21 August 2026. This section has appeared here since 13 August 2026, but nothing was recorded before 21 August 2026: our own browser security rules did not permit the required script, even where consent had been given. This was corrected in the version dated 21 August 2026; recording began only with that version.
On our public pages, we use Microsoft Clarity , a service provided by Microsoft Ireland Operations Limited. Clarity records the course of a visit—mouse movements, clicks, scrolling, screen size and the structure of the page displayed—and creates a replay of the visit and analyses of which areas of a page are used. The sole purpose is to identify where visitors encounter difficulties on our pages.
Only with your consent, and only on some pages. Without your consent, the Clarity script is not loaded and no connection to Microsoft is established. You give consent through the 'Session recording' category in our cookie settings; it is off by default and separate from the 'Analytics' and 'Marketing' categories. In the signed-in area, school area and administration, there is no recording; the same applies to the page where you set up your personal account immediately after registration, the email-verification page, pages that process a one-time verification link (password reset and email verification, recovery-address verification, parental-consent confirmation and matching a cancellation request), and individual public pages aimed directly at school pupils.
Masking. Clarity masks the contents of input fields and selection menus in every operating mode; they cannot be read in a replay. We have also set masking of other page content to the strictest level Microsoft offers. The controlling setting is held by Microsoft; we can therefore describe its effect but cannot guarantee it.
Cookies. With consent, Clarity sets exactly two cookies in our configuration. We measured this in our own project on 21 August 2026 because Microsoft's documentation does not specify their lifetimes:
The advertising cookies Microsoft generally provides for Clarity ( _uetmsclkid , _uetvid ) or cookies from Microsoft domains ( MUID , CLID , ANONCHK , MR , SM ) were not found in the same measurement. We have disabled the sharing of data collected through Clarity with Microsoft Advertising; our browser security rules also prohibit the request used for that matching. You can view the cookies that have been set in your browser settings and delete them there at any time.
Microsoft's role and transfer to the United States. According to Microsoft, it acts as a controller for Clarity, rather than as our processor; Microsoft has access to the collected data and processes it for its own purposes. Data is stored in the Microsoft Azure cloud; Microsoft does not guarantee that Clarity is restricted to data centres in the European Union. Data is transferred to Microsoft Corporation in the United States of America on the basis of the European Commission's standard contractual clauses; Microsoft Corporation is also certified under the EU-US Data Privacy Framework. Nevertheless, a level of protection fully equivalent to European standards cannot be guaranteed. The legal basis for storing and reading information on your device is section 25(1) TDDDG; for subsequent processing and transfer, it is Article 6(1)(a) and Article 49(1)(a) GDPR (consent).
Withdrawing consent. You can withdraw your consent at any time with effect for the future. To do so, open the cookie settings and disable the 'Session recording' category or select 'Necessary only'. The settings are available in our website's footer and, when signed in, also in your profile under 'Data & Privacy'. Any recording in progress then stops immediately and does not start again on a later visit. The lawfulness of processing before withdrawal is unaffected.
After checkout is completed, your browser creates an entry in local storage. It consists of a fixed name and the checkout number ( quanta.purchase.v1:<number> ) and the timestamp. Its sole purpose is to avoid counting the same completed purchase twice: the completion page may be opened more than once during a checkout, for example when returning from the payment page or reloading.
For payment methods whose receipt is confirmed later (such as SEPA direct debit or iDEAL), two such entries are created: a first one with the suffix ( :pending ) when checkout is completed, and a second once payment has been received. Both expire after the same period.
This entry is created independently of your consent because it limits measurement rather than enabling it. It contains no account data, payment data or identifier extending beyond that one checkout, and is not transmitted to anyone. After 6 hours it ceases to have effect; it is not deleted at that point, and you can remove it at any time through your browser settings. If local storage is unavailable, for example in private browsing, nothing is stored and deduplication works only within the open page.
The legal basis for storage is section 25(2) no. 2 TDDDG (strictly necessary to provide the service expressly requested), and for subsequent processing Article 6(1)(f) GDPR (legitimate interest in accurate measurement of completed purchases without double counting).
This website is provided through Firebase App Hosting (Google LLC, 1600 Amphitheatre Parkway, Mountain View, CA 94043, USA). The application runs in the region europe-west4 (the Netherlands) . Serverless background functions run in europe-west1 (Belgium). Google processes access data (IP address and timestamp) in accordance with the Firebase Privacy Policy . Google LLC is certified under the EU-US Data Privacy Framework.
You have the following rights against us in relation to your personal data:
To exercise your rights, please contact: amos@am-creativetech.com
Without prejudice to any other administrative or judicial remedy, you have the right to lodge a complaint with a supervisory authority. The competent supervisory authority is the Saxon Data Protection Commissioner.
During your visit to the website, we use the widely adopted SSL (Secure Socket Layer) protocol together with the highest encryption level supported by your browser. All personal data is transmitted in encrypted form. Passwords are stored only in hashed and salted form using industry-standard algorithms.
This Privacy Policy is currently valid and is dated October 2026. Further development of our website or changes in legal requirements may make it necessary to amend this Privacy Policy.
To ensure the service's technical and economic viability, Quanta collects internal usage statistics based on the signed-in user account. This includes:
The legal basis for this processing is Article 6(1)(b) GDPR (performance of a contract) and Article 6(1)(f) GDPR (the provider's legitimate interests in cost control and service optimisation). The data is used exclusively internally and is not disclosed to third parties for advertising purposes.
Users can voluntarily publish flashcard sets and their topics' multiple-choice tests in the community ( Quanta Community Decks ). The following applies:
For requests to delete content from public decks, please contact: amos@am-creativetech.com .
Educational institutions (schools) can use Quanta as institutional customers for their teachers and pupils. In this situation, the following data-protection provisions apply:
The legal basis for processing by Quanta is the data processing agreement in conjunction with Article 28 GDPR. Please address requests for access, rectification or erasure concerning school-managed data to your school as controller.
AI-generated learning content is primarily supported by verified, openly licensed knowledge sources. For every card supported in this way, the app identifies the source used. The content sources integrated include:
For sources under a Share-Alike licence (CC BY-SA), the Share-Alike principle applies to content derived from them; the respective source and licence are identified. No card is created without a source: content that cannot be supported by a source used in its creation is discarded and not displayed (section 15(2) of the Terms of Use). Content generated before the switch to this process continues to carry an 'AI knowledge' label in the app, identifying its ungrounded origin.
For quality assurance and further development of AI generation, Quanta maintains an internal library of quality-checked, AI-generated learning content. Only flashcards that have passed the quotation-evidence check against an openly licensed source and whose sources all permit adaptation are added to this archive (see section 15).
Legal basis: Article 6(1)(f) GDPR (the provider's legitimate interest in quality assurance, error analysis and further development of AI generation). You have the right to object to processing under Article 21 GDPR. To do so, please contact amos@am-creativetech.com .
To provide the service, we use the following processors and subprocessors and, identified separately in each case, a joint controller under Article 26 GDPR and independent controllers. This overview lists all recipients described in the individual sections above:
| Service | Purpose | Region / transfer |
|---|---|---|
| Google / Firebase (Google LLC) | Authentication, database (Firestore), hosting and Cloud Functions; for the podcast function, text-to-speech (Cloud Text-to-Speech), Hebrew speech recognition (Cloud Speech-to-Text) and audio-file processing, section 4e | Firebase Authentication: USA. Firestore and database backups: europe-west3 (Frankfurt). File storage and Functions: europe-west1 (Belgium). App Hosting: europe-west4 (the Netherlands). Text-to-speech: EU endpoint eu-texttospeech.googleapis.com. Hebrew speech recognition and audio-file processing: europe-west4 (the Netherlands). Third-country transfers: EU-US Data Privacy Framework and applicable data-protection terms. |
| Mistral AI SAS (15 rue des Halles, 75001 Paris, France) | AI generation from learning content, including podcast scripts; automated reading of uploaded PDF documents and photos; transcription of voice recordings (dictation and podcast recordings in German and English); learning tutor and exam simulation. No training with your content; no additional core account data is automatically included. Personal information in your user content may be transmitted along with it. For covered stateless requests, the provider setting restricting storage to the processing period is activated. Files and batch processing are excluded and are deleted separately. | Based in France. Text processing uses the provider's regional EU endpoint (inference takes place in the EU or EFTA according to its documentation; control and administrative data may be processed elsewhere); file reading and transcription use the standard endpoint, for which the provider does not guarantee a specific processing location. According to the provider (as of 18 August 2026), processing infrastructure is predominantly in the EU and EEA, with some of its service providers also outside the EU; standard contractual clauses under Implementing Decision (EU) 2021/914. |
| Google reCAPTCHA Enterprise / Firebase App Check (Google LLC) | Protection against automated misuse in the signed-in area (risk assessment based on interaction and device signals and IP address), section 6c | USA; EU-US Data Privacy Framework |
| Google Analytics 4 (Google Ireland Limited) | Web analytics, only with consent (Consent Mode v2) | EU-US Data Privacy Framework |
| Microsoft Ireland Operations Limited (Microsoft Clarity) | Session recording only on public pages and only with consent to the 'Session recording' category; not used in the signed-in area, school area or administration. According to Microsoft, it acts as an independent controller, not as our processor, section 6e | Ireland, with transfer to Microsoft Corporation in the USA (standard contractual clauses, EU-US Data Privacy Framework) |
| Umami (Umami Software, Inc.) | Cookieless, anonymous audience measurement of public pages (no cookies, no personal data) | USA |
| Meta Platforms Ireland Limited | Advertising-campaign measurement (Meta Pixel in the browser and server-side reporting of completed paid purchases through the Conversions API), only with consent to the Marketing category; joint controllership under Article 26 GDPR, not a processor | Ireland, with transfer to the USA (EU-US Data Privacy Framework, standard contractual clauses) |
| Stripe, Inc. | Payment processing (paid plans) | EU-US Data Privacy Framework |
| Upstash, Inc. | Rate limiting / misuse prevention (Redis) | EU data centre (Frankfurt) |
| Resend, Inc. | Transactional delivery of account, security and consent emails, and of contract and cancellation confirmations (sections 5, 5a and 6g) | USA |
| Google Workspace (Google Ireland Limited) | Email inbox: messages to us, including entry without purchase in the first-year student promotion prize draw and the selected person's reply containing their name and shipping address, section 5b; processor | EU, with possible transfer to Google LLC in the USA (EU-US Data Privacy Framework) |
| Parcel or postal service | Delivery of the first-year student promotion prize; only the selected person's name and shipping address; independent controller, not a processor, section 5b | Germany |
Version 3.4 , as of 16 October 2026