Privacy policy
Legal
This page explains what data KI können processes about you, for what purpose, on what legal basis, for how long and who receives it.
The key points
- The controller is Cüneyt Hizli, Singen, Germany (section 1).
- Account: All you need is an email address. There is no password, and we do not ask for your name for the account.
- Learning: We store your learning profile, your answers, your progress, your instructions, your check results and your own projects so that the app works for you.
- Trial lesson without an account: What you do there stays on your device only, until you take it over into your account yourself after signing in (section 5).
- Your own AI: You carry out the instruction in an AI of your choice. What you enter there goes to its provider, not to us (section 4.11).
- Cookies: We only use technically necessary cookies. Public pages set no cookies as long as you do not save a theme there. We do not embed advertising or analytics services of other providers.
- Location: Our server is in the EU, at Hetzner in Helsinki, Finland. For hosting, we do not transfer any data to countries outside the EU.
- AI: For the AI features we send texts to the AI provider Anthropic, for example your instruction when you have it evaluated. Anthropic stores the data in the USA and may not use it to train its AI models. Please do not enter names of customers, patients or colleagues.
- Payment: The course and the AI extension are sold by Paddle. Paddle processes your payment data under its own responsibility. From Paddle we only receive identifiers, the amount, the currency, the country and the status of your purchase, no payment data.
- Export and deletion: In your profile under Settings › “Data and account” you download all your data as a ZIP file or delete your account. It is then blocked immediately and finally deleted after 30 days (section 8).
- Your rights: access, rectification, erasure, restriction, data portability, objection and complaint to a supervisory authority (sections 11 and 12).
1. Controller
The controller responsible for processing your personal data in KI können (Art. 4(7) GDPR) is:
- Name: Cüneyt Hizli
- Address: Freibühlstraße 15, 78224 Singen, Germany
- Email: kontakt@kikoennen.de
“We” and “us” in this policy mean the controller. Further details are in the imprint.
2. What this policy covers
This policy applies to
- the website and the web app KI können at kikoennen.de,
- the address sandbox.kikoennen.de, where the app shows the example tools, and
- the emails we send you.
It does not apply to websites of other providers, even if we link to them, for example Paddle’s customer portal or AI services such as ChatGPT and Claude that you open from the app (section 4.11). Their own privacy policies apply there.
3. Please don’t enter other people’s data
In some places you write freely: in the learning profile (for example job, tasks, interests) and in your instructions. We store what you write there. Some of these texts go to the AI provider (section 4.9).
- Please do not enter names of customers, patients or colleagues, nor any other details by which people can be identified.
- Do not enter health data or other specially protected data (Art. 9 GDPR), neither about yourself nor about others. We do not ask for it.
- If the app finds an email address or phone number in your instruction, it points this out to you. The app does not recognize names.
- The same applies when you copy an instruction from the app into your own AI: data of customers or patients belongs only in an AI that your business has approved for this, or in a service with a data processing agreement (section 4.11).
4. What data we process and why
For each processing activity, this section states which data we process, for what purpose, on what legal basis and for how long. Section 8 summarizes the retention periods.
4.1 Visiting the website
When you open a page, your browser automatically sends data to our server:
- your IP address,
- the requested address with its parameters,
- date and time,
- details of your browser, for example the browser identifier (user agent), preferred language and the previously visited page (referrer),
- our cookies, if set (section 5).
Purpose: delivering pages, redirecting addresses without a language (for example “/”) to your language, and protecting the signed-in area.
Legal basis: Art. 6(1)(f) GDPR. Our legitimate interest is a secure and working website. In the signed-in area also Art. 6(1)(b) GDPR.
Retention: The app itself does not store page views. Our web server writes an access log: time, requested address, response code, size and duration, browser identifier and the shortened IP address (for IPv4 the first three blocks, for IPv6 the first 48 bits). It blacks out secret keys in addresses (for example from sign-in links) and does not record cookies or the previously visited page (referrer). The log is started anew every day and deleted after 13 days.
No content from other providers: We deliver fonts, images and scripts from our own server. When you visit our pages, your browser loads nothing from servers of other providers, so for example no fonts from Google and no videos from YouTube. Payment is an exception: only when you click “Continue to payment at Paddle” in the order summary does your browser load Paddle’s payment form (section 4.15).
Pages for reading: You read the guides and the FAQ page without an account. They have no forms and, like all public pages, set no cookies as long as you do not save a theme (section 5). The app does not store which articles or questions you read; only the access log above applies. Links to the sources of a guide open in a new tab, and the linked page does not learn which of our pages you came from.
Preview images for shared links: For the start page and every guide, our server creates a preview image with text and the logo, for guides with their drawn title picture. If you share a link to our pages, for example in a messenger or a social network, that service fetches the page and the preview image from us with its own server. We learn nothing about you in the process; what the service processes about you is governed by its privacy policy.
4.2 Error and operations logs
Our software writes technical logs. They contain the time, the event and cleaned-up error messages; for failed requests also the type of request and the path without parameters; for AI requests the function, result, duration, cost, model, prompt version and the identifier of the request at Anthropic.
The logs deliberately contain no IP addresses, email addresses, sign-in links, license keys or cookies and no texts of yours. Two technical filters remove such details. The lines that the web framework Next.js itself writes about server errors also pass through the same filter.
For some actions on your account, the log contains the internal identifier of your account (a random identifier, not your email address): the data export, deleting the account and the final deletion, the email “Your account is deactivated”, the license key and downloading the offline edition (then also the purchase identifier), and changes in the admin area. This lets us trace errors and misuse in these actions.
Purpose: finding errors and securing operations.
Legal basis: Art. 6(1)(f) GDPR. Our legitimate interest is stable and secure operation.
Retention: The logs are rotated by size: at most about 50 MB remain per service, and older entries are overwritten. There is therefore no fixed period in days.
4.3 Protection against abuse
So that nobody abuses the sign-in or the app, we count requests:
- Per IP address at sign-in, for example at most 5 sign-in links in 15 minutes. For this we store the IP address (for IPv6 only the first 64 bits), the requested path, a counter and the time.
- Per email address at most 3 sign-in emails in 15 minutes. For this we do not store the address but a check value we calculate from it with a secret key (HMAC).
- Per account answers, instructions and profile changes per minute and per day; for the AI features also job matches (at most 3 matches by AI in 7 days); for purchases, starting the payment and checking whether the payment has arrived. For the data export at most 3 exports and for deleting the account at most 5 attempts per hour, plus at most 3 links for signing in again per hour (section 4.21). For the license key at most 20 retrievals and for the offline edition at most 5 downloads per hour (section 4.16). For this we store a key with your account identifier, a counter and the start of the time window.
- For the “Cancel contracts here” and “Withdraw from contract here” forms: at most 10 declarations per internet connection in 10 minutes; at most 5 confirmations per day to the same email address; if you send them signed in to your account’s address, at most 5 per day for your account. For this we store neither the IP address nor the email address, only check values we calculate from them with a secret key (HMAC), with a counter and the time. In addition, we count for the whole app how many confirmations for declarations without sign-in are sent in one hour (at most 30); we send any further ones a little later. If a limit is reached, the page shows you other ways, for example email. Above the limit per email address or account we still store a withdrawal; only the immediate confirmation by email is not sent.
- For example tools and the trial lesson without an account: Your browser loads an example tool (in the trial lesson, with or without an account, and with the course in the workshop) from our sandbox server (its own address sandbox.kikoennen.de, without a cookie). Our server processes your IP address as with every page view (section 4.1) and counts the retrievals per address (for IPv6 per /64 network), at most 60 per minute. This counter is held only in the server’s memory, not in the database, and applies for one minute. If you confirm after signing in that your trial lesson should be taken over into your account, we count at most 60 takeovers per hour per internet connection (so it also works for a group behind a shared connection, for example at a school) and at most 10 per hour per account. For this we store neither the IP address nor the account identifier, only check values we calculate from them with a secret key (HMAC), with a counter and the time.
Purpose: preventing abuse, for example mass sign-in attempts, guessing links, flooding a mailbox or overloading the app.
Legal basis: Art. 6(1)(f) GDPR. Our legitimate interest is the security and availability of the app.
Retention: A daily cleanup deletes entries older than 24 hours. They are therefore stored for about 24 to 48 hours.
4.4 Account and sign-in by email link
You sign in without a password. You enter your email address, and we send you a sign-in link. It is valid for 15 minutes and only once. With the first link you use, we create your account. If you do not use the link, no account is created.
Data with the request: your email address, the language of the page and, where applicable, a job or field that was preselected in the link you came through.
Data in the account: an internal identifier, your email address and whether it is confirmed, your language, your time zone (section 4.6), your role (normal account or admin account) and times such as creation and last change.
Agreement to the terms: Directly above the “Send sign-in link” button it says that by clicking you accept the terms. When you create your account with the link, we store which version you agreed to and when (type of agreement, version, time), so that we can prove which terms apply to your contract. We store this again with every purchase (section 4.15).
We do not ask for your name for the account. There are no passwords and no sign-in through other services such as Google.
Purpose: providing your account, signing you in and sending you emails in your language.
Legal basis: Art. 6(1)(b) GDPR (contract for the use of KI können); for proof of agreement to the terms also Art. 6(1)(f) GDPR (legitimate interest: being able to prove which terms apply).
Retention:
- Sign-in request: We store the link only as a check value (hash), with your email address next to it in plain text. If you use the link, we delete both immediately. Otherwise the daily cleanup deletes the entry after it expires, at the latest after about 24 hours.
- Account: until deletion (section 8).
- Agreement to the terms: as long as the account exists; with a purchase also in the order process (section 4.15).
4.5 Staying signed in (session)
After you sign in, we set a session cookie (section 5). In our database we store for the session:
- a random session identifier,
- the IP address and the browser identifier (user agent) at sign-in,
- your account identifier and the expiry date.
The session is valid for 30 days. When you use the app, it is extended, at most once a day. It ends when you sign out.
The software we use for sign-in stores the IP address and browser identifier by default. We do not analyze them. They serve to trace, in case of abuse or a security incident, where a session was started from.
Legal basis: for the session Art. 6(1)(b) GDPR; for IP address and browser identifier Art. 6(1)(f) GDPR. Our legitimate interest is the security of the accounts.
Retention: until you sign out. Otherwise the session ends 30 days after the last extension, and the daily cleanup deletes it. When an account is deleted, we immediately delete all its sessions.
4.6 Learning profile
After your first sign-in, the app asks for your learning profile in three steps (field, AI experience, learning time); there you can optionally state your job more precisely. It offers further questions (job, work situation, tasks, time wasters, goals, interests) later under “Complete your profile”; they are optional. You can change all details later in your profile.
Data:
- your job as free text (up to 120 characters); if it matches our job catalog, also the catalog job and the field;
- your work situation (employed, self-employed, founding, in training, other), when founding also the area of the business;
- typical tasks and time wasters (selection and free text);
- your goals, your AI experience (four levels), your interests (selection and free text) and your learning time per day;
- on how many days per week you want to learn (“Learning rhythm”);
- your preferred design and which step you are at;
- your chosen path in modules that offer several paths (currently only “Set up a tool and get started”: which AI tool you set up).
If you come through a link that specifies a job or field, we take this selection into your profile if nothing is there yet.
You do not choose the path during setup but in the lesson “Setup step by step”. The choice is optional; without it the lessons show a general version.
Instruction library: your catalog job decides which instructions of a profession page you get in the Instruction library with the free version: those of your job’s page, if there is one (terms 4.1). The library shows these instructions first. With the course it also sets the filter “Profession” to your job, otherwise to your field or to all jobs; you can change it at any time. We store nothing extra for this.
Time zone: In every step of the setup your browser sends the time zone (for example “Europe/Berlin”), even if you skip a step. Without JavaScript we use “Europe/Berlin”. We need the time zone so that daily quotas restart at midnight your local time, reviews fall due on the right day and the month for the AI cost limit counts correctly (section 4.13). The time zone is not shown in the settings.
Purpose: putting together your learning path, suggesting a design, selecting examples and comparisons for your job and your field (section 4.9 c), showing you the steps for your chosen path in the lessons and determining the free instructions and the preselected filter in the Instruction library.
Legal basis: Art. 6(1)(b) GDPR. Learning tailored to you is the core of our service.
Retention: until your account is deleted. Changes overwrite the old details. If your learning path changes as a result, we store it as a new version; older versions remain stored.
4.7 Learning progress and exercises
Data:
- your learning path with its versions;
- per lesson the status, result (0 to 100) and the times of start and completion. A lesson starts with your first answer; we do not record mere reading;
- every answer to questions, including reviews, the second try at the end of a lesson (“One more time”), the module check at the end of a module, the mixed practice round (“Practice”) and the “Quick quiz”: chosen answers, orders or words, whether the answer was correct, and the time. Whether a lesson counts as done (at least half of its quiz questions right once) is calculated by the app from these answers; we store nothing extra for this. From your answers in the “Quick quiz” the app picks the questions of your next rounds (those you last answered wrongly are preferred); these answers count toward your learning days, but not toward the status of a lesson, the review or your level in the skills;
- whether and when you skipped the instruction in the Baukasten of a lesson (“Skip the instruction”);
- your review schedule (after 1, 3 and 7 days) and your level in six skills;
- your written instructions from exercises and writing questions (up to 4,000 characters) with your self-check or the AI evaluation (section 4.9). Every version remains stored so that you can compare versions. The app calculates the rule-based feedback on the Baukasten without AI from the building blocks of your instruction;
- your check result at the end of a lesson or of your own project (“Check the result”), per version of the instruction: the ticked items of the checklist, whether it worked (yes, partly, no), in which AI service you tried it (only the identifier from our list, for example “chatgpt” or “other”), your correction instruction (up to 4,000 characters) and your note (up to 500 characters), with times;
- per module the highest state of your mascot (how many building-block groups of the phoenix light up and whether it has shone once), so that it never falls back.
In addition there are your answers in the placement test, your badges (for example “First result” for the first “Did it work? Yes”) and your learning days with learning minutes and whether you reached your daily goal; from these the app shows the daily ring and your weekly rhythm. We do not store the results of the mini-games; only your browser remembers the highest level reached (section 5). Whether and which cheat sheets you open, print or save as a PDF is not stored in your account; your browser does the printing and saving as a PDF on your device. In the Instruction library, too, we do not store what you search for, which filters you choose or which instructions you copy: the library loads the instructions you may use once with the page; search and filters then run in your browser, and your search terms are not sent to our server.
Purpose: showing you your progress, planning reviews, putting together the rounds of the “Quick quiz” and evaluating exercises.
Legal basis: Art. 6(1)(b) GDPR.
Retention: until your account is deleted.
4.8 Appearance and language
We store your appearance settings (theme, font size, density, “High contrast”, “Reduce motion” and color mode) and your language (German or English). With an account they are stored in your account and also in cookies on your device (section 5). Without an account we store only the theme you choose at the bottom of the page, and only in the cookie. Unless you choose otherwise, the app shows everything dark; with an account you choose “light” and “like the device” under Settings › Appearance. Whether sounds and music are on is remembered only by your device (section 5).
Purpose: showing the app the way you set it up, also after signing out on this device.
Legal basis: Art. 6(1)(b) GDPR; without an account Art. 6(1)(f) GDPR (legitimate interest: the appearance you chose).
Retention: in the account until it is deleted; cookies see section 5.
4.9 AI features
Some features use AI. For this we call the interface (API) of the AI provider Anthropic. Anthropic processes the data on our behalf (processor under Art. 28 GDPR). Our contracting partner is Anthropic Ireland, Limited, 6th Floor, South Bank House, Barrow Street, Dublin 4, D04 TR29, Ireland. More about the AI features, the models and labeling is in our notes on AI.
For all AI features:
- Server to server: The requests go from our server to Anthropic. Anthropic therefore does not receive your IP address, and we send neither your account’s email address nor a name. What you write into a text yourself is sent as described below.
- Pseudonymous identifier: If a request runs for a specific person, we send an identifier. We calculate it from your account identifier with a secret key. Anthropic can use it to recognize requests from the same person in order to prevent abuse, but cannot determine who you are.
- No training: Anthropic may not use the content to train AI models. This is stated in Anthropic’s terms for business customers (Commercial Terms of Service): Anthropic may not train models on customer content from its services.
- Location: Anthropic stores the data in the USA. The computation itself may also take place in other countries. On the transfer see section 7.
- Retention at Anthropic: According to Anthropic, API inputs and outputs are deleted within 30 days. Exceptions: if Anthropic’s automated review classifies content as a violation of Anthropic’s usage policy, Anthropic keeps the content for up to 2 years and the classification for up to 7 years; and as far as a law requires it.
- Placeholders: When matching your job (a) and when evaluating (b), the app replaces email addresses and phone numbers with placeholders before sending the text. The app never recognizes or replaces names.
- Not sent: We do not send your free texts on tasks, time wasters and interests to Anthropic.
- Background tasks: Matching your job (a) runs in the background. For this we store your account identifier in a queue, never your texts. We delete these entries 2 days after completion.
Legal basis: Art. 6(1)(b) GDPR. The AI features are part of the service you use.
a) Matching your job. If the app cannot reliably match your job to a catalog job (it tries this first without AI), it asks the AI, at most three times in 7 days per account; after that it matches by fixed rules.
- Sent: your job text (with placeholders), the language of the app and our job catalog.
- Result: a general job title, a field and, if possible, a catalog job. We store this in your profile.
- If your job is new, we create a suggestion with the general job title for our catalog. We review it and add it if appropriate.
- We also store results without reference to your account in a cache, so that the same details are not sent again.
b) Evaluating instructions. When you tap “Have it evaluated” in an exercise or submit your text in a question in which you improve an instruction, the AI evaluates it (within your quota: in the AI period; in the free version once, up to 3 times, in the trial lesson; without an account never). In the Baukasten this is possible only if your instruction contains words of your own.
- Sent: your text (up to 4,000 characters, with placeholders); a short profile of job title (from the catalog or the general title, never your free text), field, work situation and AI experience; also the task, evaluation criteria and model solution.
- Result: a level per criterion, quotes from your text as evidence, hints and a suggestion for improvement. We store the evaluation with your instruction (section 4.7). If you send the same instruction again unchanged, we show the stored evaluation without sending it again.
c) Examples for jobs. The AI does not adapt examples while you learn. We have the AI create examples, exercise tasks and comparisons for the jobs in our catalog in advance. Automated checks run over every result; after that we review the results by sampling and approve them before they appear in the app.
- Sent: only lesson texts and our catalog details on the job (for example typical tasks in this job). These requests do not run for a specific person; we do not send details about you or your pseudonymous identifier for them.
- Result: an example, values for the task and a comparison per lesson, job and language. They are content of the app without reference to persons; everyone with the same job sees the same ones.
- On our side: With the course and in the trial lesson, the app selects the approved example by the catalog job in your learning profile and the language (section 4.6). If there is none, for example because your job is not in the catalog, the lesson shows the general example and, where there is one, the comparison for your field. Nothing goes to Anthropic for this.
d) Pre-made example tools and “Before and after”. We have example tools and an AI’s answers to a weak and a good instruction created in advance, once for everyone. Only instructions and texts we have written and details from our job list are sent; these requests do not run for a specific person, and nothing about you goes to Anthropic. Automated checks run over every result; after that we review the results by sampling and approve them. They are then content of the app without reference to persons. Which examples you see is selected by the app without AI according to your job, your field and your idea in the trial lesson (section 4.6).
4.10 Workshop
The workshop is where your learning results are kept. The app does not build or store tools for you; they are created in your own AI (section 4.11).
- Lesson results: your instructions from the lessons with their versions and check results (section 4.7). They do not count against a limit.
- Own projects: name, goal, your instructions for it with all versions, evaluations (in the AI period) and check results; in the free version up to 3, with the course up to 50.
- Example tools (course): the example tools for your job and your field (section 4.9 d). They open on our sandbox server.
- Isolated examples: Example tools run under their own address (sandbox.kikoennen.de) in an isolated frame without access to the app, without internet access and without browser storage. Our cookies are not sent to this address. What you type into an example stays in the memory of your browser tab; neither we nor anyone else receives it, and it is gone when you close it. Examples contain no data about you, so we deliver them without a signature and let them be cached briefly; we only count the number of retrievals per address (section 4.3).
Purpose: storing and showing you your learning results and projects; showing examples.
Legal basis: Art. 6(1)(b) GDPR.
Retention: until you delete an own project or your account is deleted. We delete the check results of a project together with the project. Individual lesson results cannot be deleted; they remain until your account is deleted.
4.11 Your own AI
At the end of a lesson you carry out your instruction in an AI of your choice (“Use it in your AI”).
- Copying and opening: “Copy instruction” puts the instruction on your device’s clipboard. “Open in ChatGPT” and “Open in Claude” also copy it and then open the service’s start page in a new tab, without writing the instruction into the address and without partner, campaign or tracking parameters. The new tab does not learn which page it was opened from. “Other or company AI” only copies.
- What the provider receives: What you paste or type into the service goes from your device to its provider, for example OpenAI or Anthropic, under its terms and privacy policy. We transfer nothing for this and receive nothing back from the provider. Whether the provider uses your inputs for training and how you switch that off is up to the provider; the lessons and the AI finder state the current position with a source.
- What we store: We only learn what comes out of it if you enter it yourself under “Check the result” (section 4.7).
Legal basis: Art. 6(1)(b) GDPR for copying and opening as part of the lesson.
4.12 AI finder
The AI finder (publicly on the “AI finder” page and in the app in the workshop) asks about your purpose, a requirement of your company, data of customers or patients, budget and device, and recommends AI services by fixed rules, without AI.
- Your answers are only in the page address. The app stores them neither in the database nor in the browser; with the course only the identifier of your own project is added if you open the finder from a project.
- Like every requested address, it appears with the shortened IP address in our web server’s access log and is deleted after 13 days (section 4.1). The web server blacks out the answers there (company AI, data of customers or patients, budget, device); only the purpose remains visible.
Legal basis: Art. 6(1)(f) GDPR (legitimate interest: a recommendation without an account), in the app Art. 6(1)(b) GDPR.
4.13 Quotas and AI costs
There are quotas for the AI features. For this we store:
- per day and function, how often you used it. These daily counters also show how much you used in a month in the AI period (monthly limits);
- one-time trial uses in the free version;
- per AI request: function, model, amount of text units processed (tokens), estimated cost, result, time and whether the request was counted against a paid quota, linked to your account.
Purpose: keeping quotas, limiting costs (with a monthly cost limit per account, separate for free, the first 12 months and the AI extension, and a shared daily budget against which use in the AI period within the quotas is not counted; see the terms), preventing abuse and analyzing costs.
Legal basis: Art. 6(1)(b) GDPR, because quotas are part of the service, and Art. 6(1)(f) GDPR. Our legitimate interest is controlling costs and preventing abuse.
Retention: The counters until your account is deleted. After that, the cost rows remain without reference to you for cost analyses.
4.14 Emails
- Sign-in email: with your sign-in link, in your language.
- Confirmations of cancellation and withdrawal: to the email address you enter in the form (section 4.17).
- Contract confirmation: after every purchase, to your account’s address, with the services and quotas, price, payment, the end of the AI period, the version of the terms, the details of the provider, the withdrawal information with the model withdrawal form and, if you requested it, the immediate start (§ 312f BGB).
- Messages about your account: a new sign-in link for confirmation before export or deletion and, after deleting, the email “Your account is deactivated” with the day of the final deletion (section 4.21).
We use your email address for sign-in and for messages about your account and your purchase. We do not send you advertising. Our emails contain no tracking pixels and load no content from other servers.
Sending: through Lettermint B.V., Willemsvaart 16 B, 8019 AB Zwolle, Netherlands, as our processor (data processing agreement under Art. 28 GDPR as part of Lettermint’s terms of use). According to Lettermint, it processes the emails only in the EU; according to Lettermint, its subprocessors are located in the Netherlands, in Germany and in Europe. Open and click tracking is switched off at Lettermint. Your email provider receives the email.
Legal basis: Art. 6(1)(b) GDPR; for confirmations required by law Art. 6(1)(c) GDPR.
Retention: We do not store sent emails ourselves; what a confirmation contains is stored with the declaration or the purchase (sections 4.15 and 4.17). At the sending service: content, recipient address and delivery events 28 days after sending, events for analyses at most 90 days, backups up to 14 days longer (Lettermint’s data processing agreement).
4.15 Purchase and payment through Paddle
The course and the AI extension are sold by Paddle.com Market Limited, 30 Old Bailey, London EC4M 7AU, United Kingdom (“Paddle”). Paddle is the reseller (merchant of record) and handles payment, invoice, sales tax and refunds.
- Paddle is itself responsible for the data you enter when buying, for example email address, country and payment details. Paddle and we are each independent controllers. We never see your payment data, such as your card number. How Paddle handles your data is described in Paddle’s privacy notice: paddle.com/legal/privacy.
- Order summary: Before you go to Paddle, the app shows you a summary. If you click “Continue to payment at Paddle” there, we create an order process. It contains your account, the product (course or AI extension), the price identifier at Paddle, the payment provider, the version of the terms you accept by doing so, the language and, if you request the immediate start, the time and version of that wording. Once the order is completed, the time and the identifier of the payment at Paddle are added. Your email address is not in it. We also store your agreement to the terms as at sign-in (section 4.4).
- Payment page: Only after this click does your browser load a script from Paddle (Paddle.js) from Paddle’s server, and Paddle’s payment form opens over our page. The same applies to our payment page that Paddle may link to in its emails, for example for a payment. On all other pages the app loads nothing from Paddle.
- What we give to Paddle: When the payment form opens, the price identifier of the product, your account’s email address for prefilling, the language of the page and the identifier of the order process. We use this identifier to assign the payment to your account. When we query the status of a payment, we send its identifier. If you open Paddle’s customer portal on the “Your purchase” page, for example for your invoices, we send Paddle your customer identifier at Paddle. What goes to Paddle with a withdrawal is described in section 4.17. Paddle does not receive your learning profile, your instructions or your progress.
- What we receive from Paddle: Paddle notifies us of the payment and later changes with messages (webhooks). After payment we also query the status at Paddle, and a daily check reconciles purchases with Paddle. Of this we take over only: the identifier of the payment, your customer identifier at Paddle, the status, the price identifier and the quantity, the total amount, the currency, the time of completion, the identifier of our order process and, of the address you enter at Paddle, only the country (as a country code, for example “DE”). We store the country with your purchase so that we can analyze in which countries purchases are made (for target markets and prices per country). For refunds and chargebacks of your payment – those we request from Paddle, for example after a withdrawal, and those we trigger manually in Paddle or that Paddle reports – we take over only the identifier, type, time and status and the identifier of the payment. The daily check also reads the amounts in order to add up partial refunds; we do not store them. Of other transactions on your payment, such as credits, we store only the identifier of the message. We do not take over names, email addresses, the rest of the address or payment details. Of every message we store only a selection without amounts (at most 16 KB) with the type, time and result of processing, separately from your account.
- What we store per purchase: payment provider, identifier of the payment, your customer identifier at Paddle, the product, the status, the price identifier, the amount, the currency, the time of purchase, the time the purchase ended, the country (country code), the quotas that applied at purchase and that you keep at least, and whether it is a test purchase; plus the link to the order process, when we sent the contract confirmation and when we first handed out the license key (section 4.16).
- Access: What you can use follows from these entries. Your access to a purchase ends when Paddle approves a full refund, when the payment is charged back and when Paddle reports a chargeback warning (according to Paddle, the amount has then already been refunded). If Paddle wins the dispute, access is restored. For this we only record the status and end of the purchase.
- Immediate start: If you request at purchase that we start the service immediately (see the withdrawal information), we store this with the time and version of the wording in the order process. It also appears in your contract confirmation (section 4.14).
- Notices to us: If the app cannot assign a payment to an account, an account has two courses, a payment contains an unusual quantity or there is a partial refund or chargeback for a purchase, we receive an email with the identifiers at Paddle and, with two courses, the account identifier, so that we can check and refund if necessary.
- Invoices and payments are shown in Paddle’s customer portal. Paddle is responsible there.
Purpose: Paddle sells and bills the course and the AI extension; we unlock them for you, keep your quotas and prove what was agreed.
Legal basis: for our processing Art. 6(1)(b) GDPR; for statutory retention obligations and the request for an immediate start (§ 312f BGB) Art. 6(1)(c) GDPR; for proof Art. 6(1)(f) GDPR (legitimate interest: being able to prove what was agreed).
Retention:
- Purchase data: until your account is finally deleted (section 8).
- Order processes: we delete uncompleted ones 30 days after creation, or earlier if the account is deleted. Completed ones remain as long as the purchase is in your account. If the account is deleted, they remain without reference to your account until the end of the third calendar year after completion, as proof of agreement and immediate start.
- Webhook messages: 13 months, even if the account is deleted earlier, because they are not tied to the account.
- Records we must keep for accounting (for example Paddle’s statements): until the end of the statutory retention period, for accounting vouchers 8 years from the end of the year (§ 147 AO).
Transfer: Paddle is located in the United Kingdom (section 7).
4.16 License key and offline edition
The course includes a license key and an offline edition (see the terms). You find both on the “Your purchase” page.
- When they are available: Only when the withdrawal period has expired and neither a withdrawal of your account nor a refund of this purchase is pending. Every time before we hand out a key or the offline edition, we ask Paddle whether there is a refund or chargeback for your payment; for this we send Paddle only the identifier of the payment.
- License key: a text signed with our key. It contains the identifier of your purchase, the product, the time of issue and who the license is for: the name of your account if it has one, otherwise your email address. The app does not ask for a name, so as a rule your email address appears there. We show it to you when you retrieve it on the “Your purchase” page.
- What we store: with your purchase, the time we first handed out the key or the offline edition. We do not store the key itself; we generate it anew with every retrieval, and it stays the same as long as the name or address does not change. It does not appear in our logs; with retrieval and download your account identifier appears there (section 4.2). To prevent abuse, we count retrievals and downloads per account (section 4.3).
- Offline edition: a ZIP file that you download. It is the same for everyone and contains no data about you. It runs in your browser from the file, without our server, without internet and without AI features. If you enter your license key there, it checks it without connecting to us and shows who the license is for.
- What the offline edition stores on your device: in your browser’s storage the license key, your progress, the language, the color mode and, if you switch off the sounds, that as well. We receive none of this. You can save your progress there as a file and load it again. You can delete it through your browser’s website data.
- What you should know: The key and the file are with you. We cannot retrieve or change them. Your name or email address remains in the key even if you delete your account. Do not pass the key on (see the terms).
Purpose: making the course accessible to you without our server and showing who the license is for; making sure before handing it out that the purchase has not been withdrawn or refunded.
Legal basis: Art. 6(1)(b) GDPR; for the query to Paddle and the counters against abuse Art. 6(1)(f) GDPR (legitimate interest: not handing out a key for a refunded purchase).
Retention: the time of handing out like the other purchase data (until your account is finally deleted). What the offline edition stores on your device stays with you only.
4.17 Canceling and withdrawing
At the bottom of every page there are the buttons “Cancel contracts here” and “Withdraw from contract here”, and in the account also on the “Your purchase” page. You can reach the forms without signing in.
- Canceling: You state whether you cancel ordinarily or extraordinarily (then with a reason), your account’s email address, which contract you cancel and as of when.
- Withdrawing: You enter your name, the contract and an email address for the confirmation and confirm with “Confirm withdrawal”. We ask for your name only here, because the law provides for it with a withdrawal.
- Assignment: When you are signed in, your details are already filled in. Without sign-in we look up the account by exactly the email address given, never a deleted one. We link a withdrawal to a purchase only if exactly one purchase of the account lies within its 14 days; otherwise we assign it by hand. We link a cancellation to the purchase within its 14 days, otherwise to the course, otherwise to the newest AI extension.
- What we store: per declaration the type (cancellation or withdrawal), your details from the form (name only with a withdrawal; with a cancellation also the type, the reason for an extraordinary cancellation and the desired end), date and time of receipt with time zone and language, whether you were signed in, account and purchase with the identifier of the payment at Paddle if we could assign them, when the confirmation was sent or, if that failed, only the type of error, and the processing status (with a refund we request from Paddle, also when we requested it, its identifier at Paddle and when we informed you of the outcome by email).
- Confirmation: We send it immediately by email to the address from the form. Above the limit per address or account (section 4.3), the immediate email is not sent for a withdrawal. The page afterwards shows your declaration and offers it as a file. It can be reached for 30 days through a link with a check value. We make links in your free texts harmless in the emails.
- What a cancellation does: It changes nothing about your purchase, refunds nothing and states no end date. We record it and read it personally. We pass nothing on to Paddle.
- What goes to Paddle with a withdrawal: only the identifier of your purchase, when we request the refund, with the reason “Withdrawal (§ 356a BGB)”. Name, email address, contract details and your other details do not go to Paddle.
- Notices to us: If a declaration needs us, for example a cancellation, a withdrawal for review or a manual refund, we send ourselves a summary email at most once a day to our own address for operations. It contains per declaration: identifier, type, reason for the notice, receipt, whether signed in, contact address, name (with a withdrawal), contract details, identifier of the purchase at Paddle, whether the immediate start was requested, reason for an extraordinary cancellation, desired end, date and status. This lets us reply to you and process the declaration. For each notice we store in the app only the identifier of the declaration, the reason and the dates.
Legal basis: Art. 6(1)(c) GDPR in conjunction with § 312k BGB (cancellation) and § 356a BGB (withdrawal); Art. 6(1)(b) GDPR for processing; Art. 6(1)(f) GDPR for proof (legitimate interest: being able to prove what was declared when).
Retention: Declarations until the end of the third calendar year after receipt, counted in German time (received on October 21, 2026: until December 31, 2029). This also applies if you delete your account earlier; we then remove the link to the account and purchase, and the identifier of the payment at Paddle remains with the declaration so that a refund remains traceable. We delete the notices to us in the app together with the declaration. We keep the summary emails in our mailbox as long as we need them for processing and for follow-up questions on the declaration. On the limits against abuse see section 4.3.
4.18 Contact, support and administration
If you write to us, we process your email address, your message and our reply in order to handle your request. Our mailbox kontakt@kikoennen.de is hosted by STRATO GmbH, Otto-Ostrowski-Straße 7, 10249 Berlin, Germany, as our processor (Art. 28 GDPR).
- Legal basis: Art. 6(1)(b) GDPR if it concerns your account or a contract; otherwise Art. 6(1)(f) GDPR (legitimate interest: answering requests).
- Retention: until your request has been dealt with; longer only if there is a statutory retention obligation.
Admin area: Only people with the admin role have access. For support they see, for example, your email address (shortened in lists), your purchase with its status and your quotas, can reset a quota and can reverse the deletion of your account at your request. We record such changes in a change log (type of change, time, identifiers of the acting person and the affected account). They also review new job suggestions (section 4.9 a). Analyses of AI costs show individual people only pseudonymously. In the cases in section 4.19, authorized people can also open your instructions, check results and projects.
- Legal basis: Art. 6(1)(b) and (f) GDPR (legitimate interest: support and secure operation).
4.19 Reports and moderation
Anyone can report content in the app to us that they consider illegal (see the terms). We also respond to specific indications of violations of the terms.
- Data: with a report, the name and email address of the reporting person (optional for reports on child sexual abuse or sexual exploitation of children), their reasons, the reported content and the affected account; also our review, our decision and its reasons and, if you object, the correspondence about it.
- Review by people: We do not read your content routinely. Only with a report, with specific indications of a violation or if you ask us to, do authorized people open your instructions, check results and projects.
- Consequences: Depending on the outcome, we issue a warning, block or delete content, block features or the account, or terminate the contract. We inform you of every measure with reasons.
- Recipients: authorities, if a law requires it, for example in case of suspicion of a criminal offense that threatens the life or safety of people (Art. 18 DSA).
Purpose: handling reports, stopping illegal content and violations of the terms, giving reasons for decisions and proving them.
Legal basis: Art. 6(1)(c) GDPR in conjunction with Art. 16 to 18 DSA, insofar as KI können is a hosting service; otherwise and for enforcing the terms Art. 6(1)(f) GDPR (legitimate interest: a safe and lawful app).
Retention: as long as we need reports and decisions as proof toward authorities or in a dispute about the decision.
4.20 Hosting and backups
- Hosting: The app runs on a server of Hetzner Online GmbH, Industriestraße 25, 91710 Gunzenhausen, Germany. The server is in the Helsinki data center in Finland, that is, in the EU. Hetzner is our processor (Art. 28 GDPR). For the Finland location, Hetzner uses Hetzner Finland Oy (Tuusula, Finland) as a subprocessor for the building and technical support. According to Hetzner, with a server location in the EU the data is processed only in the EU, including in support. All data from section 4 that we store is kept there in our database.
- Backups: Every night we back up the database.
- The database backup remains on our server for 7 days.
- An encrypted copy goes to external storage, a Storage Box of Hetzner Online GmbH in the EU. There we keep the backups of the last 7 days, one per week for 4 weeks and one per month for 3 months. We delete older ones automatically.
- Data we have deleted is therefore also gone from the backups after about three months at the latest.
- If we restore a backup (only after data loss), we apply all account deletions since the backup again. For this we keep a deletion log that contains only the account identifier and the times (section 4.21). Once a month we check whether a backup can be restored; we delete the test copy immediately afterwards.
Legal basis: as for the respective purpose in section 4; for backups Art. 6(1)(f) GDPR (legitimate interest: protection against data loss).
4.21 Data export and account deletion
- Data export: In the settings under “Data and account” you download a ZIP file with “Export data”. It contains all data about your account as a machine-readable JSON file (including all answers from lessons, reviews, module checks, “Practice” and the “Quick quiz”, learning days, check results, own projects, the state of your phoenix and, after a purchase, the time the license was handed out), a readable overview with the information under Art. 15(1) GDPR and short instructions. It does not contain check values of links and keys, counters against abuse or the shared AI cache, because they say nothing about you; only Paddle has payment data (section 4.15). The file is created at the moment you request it; we do not store it.
- Confirmation: Before export and deletion your sign-in must be fresh (at most 15 minutes old). Otherwise we send you a new sign-in link that leads back to the same page.
- Account deletion: In the settings under “Data and account” you delete your account (Delete account). What happens then is described in section 8. We also send you an email “Your account is deactivated” with the day of the final deletion. If you bought the course, the deletion ends your access. We do not refund your money because of this, and we cancel nothing for it; if you are still within the withdrawal period, withdraw first (section 4.17). The deletion page tells you this beforehand.
- Deletion log: So that a deleted account stays deleted even after a backup is restored, we store the account identifier (a random number, not an email address) and the times of the deletion request, the final deletion or the reversal. We delete these entries 120 days after the final deletion, that is, after the backups’ retention has ended.
- Reversal: During the 30 days until the final deletion, we can reverse the deletion at your request. Your purchase is then back as well.
- Requests by email: If you cannot sign in, we create the export or delete your account at your request (section 11). We hand over the export file by a secure route and then delete it on our side.
Purpose: fulfilling your rights of access, data portability and erasure; enforcing deletions in backups as well.
Legal basis: Art. 6(1)(c) GDPR in conjunction with Art. 15, 17 and 20 GDPR; for the deletion log also Art. 6(1)(f) GDPR (legitimate interest: enforcing deletions permanently).
5. Cookies and storage in the browser
Cookies are small text files that your browser stores. We only use cookies that are strictly necessary for a function you explicitly want to use. We do not need consent for this (§ 25(2) no. 2 TDDDG). We process the data in them on the basis of Art. 6(1)(b) GDPR, without an account on the basis of Art. 6(1)(f) GDPR.
With you we set at most these three cookies (a fourth only with our team, see below):
- kla.session_token (with HTTPS: __Secure-kla.session_token): keeps you signed in and contains the signed session identifier. Set after sign-in. Lifetime 30 days; extended with use at most once a day; deleted when you sign out.
- kla.ui: remembers your appearance settings. Set only when you save an appearance setting: without an account the theme via “Apply” in the footer, with an account in the settings. Lifetime 1 year from the last change; remains on the device after signing out.
- kla.locale: remembers your language for addresses without a language. Set only when you change the language while signed in. Lifetime 1 year from the last change; remains on the device after signing out.
All three cookies are “httpOnly” (scripts in the browser cannot read them), “SameSite=Lax” (they are not sent with embedded requests from other websites) and, with HTTPS, “Secure” (they are only transmitted encrypted).
For our team only: people with the admin role can exclude their own device from our own audience measurement in the admin area (it is currently switched off). For this we set the cookie kla.nocount on that device (content only “1”, “httpOnly”, “SameSite=Lax”, with HTTPS “Secure”, lifetime 1 year). Only this button creates it, never for visitors or learners.
Public pages set no cookies as long as you do not save a theme there. Automated tests check this.
Trial lesson without an account: You can do the trial lesson without an account. What you do there first stays on your device only: your quiz answers, the building blocks of your instruction (including what you write into it yourself) and your check result (with your note and a correction, if you enter them) are kept by the page in your browser’s storage (localStorage, one entry with a version number, at most 16,000 characters), so that you can continue on your next visit. It also holds the chosen language, the chosen job or field, the time of the last change and – only if you tap “Save my result” – the time of that tap. No cookie is set for this. Nothing of this entry goes to our server unless you confirm the takeover after signing in. The entry is created only once you do something in the trial lesson; after 30 days without a change the page no longer uses it and deletes it on the next visit. With “Delete from this device” you remove it yourself at any time, also when it belongs to a different job choice.
- Taking it over into your account: The entry is offered only if “Save my result” was tapped on this device at most 2 hours before. If you then open the first lesson while signed in, the app first asks whether the result should be taken over. It only says what would be taken over (number of answers, whether an instruction and a check are included and whether the check contains a note or correction), not your text. Only when you tap “Take over” does your browser send the entry to us; we check it and store it in your account like a lesson you did there (Art. 6(1)(b) GDPR). Parts that no longer fit after a change to the lesson are not taken over, and we tell you which; what you wrote yourself into an instruction that no longer fits is shown to you by the app for copying. After that the entry is deleted from the device. If your account has already stored the trial lesson, we take over nothing; your instruction from there is shown to you for copying, and the entry is also deleted. With “Don’t take over” it is deleted without being sent. If the takeover does not work at the moment (a fault or too many requests) or your account cannot accept it right now (for example, the setup is not yet completed), the entry stays on the device, and the app asks again the next time you open the lesson within the 2 hours.
- Shared devices: Without “Save my result” on this device, or later than 2 hours after it, the entry is never sent. When you sign out and after an account is deleted, we delete it on the device. If an account opens a lesson on the device and no account or a different account was noted there before (see below), we delete it as well – unless “Save my result” was tapped within the last 2 hours; then the question above applies, and “Don’t take over” deletes it.
Other storage in the browser: Apart from the trial lesson entry, we use your browser’s storage (localStorage and sessionStorage) only in these cases in the app after signing in (the music also in the trial lesson without an account):
- Waiting answers: If a lesson or a question round (review, module check, “Practice”, “Quick quiz”) cannot send an answer to our server at the moment (for example on the go without a network), it keeps the answer in your browser’s storage and sends it as soon as the connection is back or you next open the lesson or round; after that the entry is gone. It contains only the identifiers of the question and its lesson and the chosen answer, no text of yours; the name of the entry also contains your account identifier and the identifier of the lesson or the kind of round. The same applies to your check result at the end of a lesson (“Check the result”): it can also contain your own text (note and correction instruction) and stays on the device only until it has been sent.
- Who learned last: So that we recognize entries of another person on the same device, we remember there the account identifier of the person who last opened a lesson. When you sign out, after your account is deleted and when another account signs in on the device, we delete waiting entries on the device.
- Game progress: If you play in the app under “Games”, your browser remembers the highest level reached per game (only the name of the game and a number, no account identifier). Nothing of this goes to our server. When you sign out and after your account is deleted, we delete the entry on the device; until then another account on the same device sees the same game progress.
- Phoenix on the learning path: When you complete a lesson for the first time, only the current tab of your browser (sessionStorage) remembers which lesson it was: the identifier of the lesson and its module, whether the module is thereby finished, and the time. With this the phoenix flies once to the next lesson the next time you open the learning path. Nothing of this goes to our server. The learning path deletes the entry as soon as it has read it; if it is older than 30 minutes, the learning path no longer uses it and deletes it as well. If you close the tab before, the browser deletes it; when you sign out and after your account is deleted, we delete it too.
- Sounds and music: If you switch off the sounds, the app remembers this on this device. If you pause the music (on the learning path, in a lesson, in the trial lesson or in a game), the app remembers that too, for all of them. These entries contain nothing about you, do not go to our server and remain after signing out, so that it stays quiet.
Example tools cannot store anything in the browser: they run in an isolated frame without access to the browser’s storage, and before approval we check that they use no storage and no cookies (section 4.10).
“Copy instruction” and “Open in …” write your instruction to your device’s clipboard, only after you tap (section 4.11). On a cheat sheet (“Copy instruction”) and in the Instruction library (“Copy”) you copy an example instruction, also only after you tap.
What else the app reads from your browser: the time zone during setup (section 4.6).
Paying at Paddle: Only after the click on “Continue to payment at Paddle” and on our payment page for links from Paddle’s emails does your browser load Paddle.js from Paddle’s server (section 4.15). Which cookies or other storage Paddle uses for this is described by Paddle in its privacy notice.
Offline page (service worker): When you are signed in, the app installs a small program in your browser (service worker). It stores exactly eight files on your device: a “No connection” page in German and in English, two stylesheets, three fonts and the app icon. Your browser shows this page when you open a page of the app without internet. The service worker stores no pages with your data, no server responses and nothing you enter; it reads nothing about you. Public pages do not install it. It remains on the device after signing out until you delete the website data in the browser; we can remove it for all users through an emergency switch. The app can also be placed on the home screen like an app (web app manifest; it stores nothing extra for this).
6. Recipients
We only pass on your data if this is necessary for the purposes described or a law requires it. We do not sell data. We do not embed advertising networks, analytics services of other providers, content delivery networks, font or video services or error reporting services.
- Hetzner Online GmbH, with the subprocessor Hetzner Finland Oy – processor: hosting and database (section 4.20). Location: Finland (EU).
- Hetzner Online GmbH (Storage Box) – processor: encrypted backups (section 4.20); the provider cannot read them without our password. Location: EU.
- Lettermint B.V. – processor: sending our emails (section 4.14). Location: Netherlands (EU).
- STRATO GmbH – processor: our email mailbox when you write to us (section 4.18). Location: Germany (EU).
- Anthropic Ireland, Limited, with subprocessors – processor: AI features (section 4.9). Location: Ireland; storage in the USA.
- Paddle.com Market Limited – independent controller: sale and payment of the course and the AI extension (section 4.15); refund after a withdrawal, for which only the identifier of the purchase is sent (section 4.17). Location: United Kingdom.
In addition:
- Authorities and courts, if we are legally obliged to (Art. 6(1)(c) GDPR).
- Your email provider receives our emails to you.
Anthropic’s subprocessors are listed at anthropic.com/subprocessors.
7. Transfers to countries outside the EU
Hosting and email mailbox: The server, backups and our mailbox are in the EU (sections 4.18 and 4.20). For these we do not transfer any data to countries outside the EU.
Anthropic (USA): Our contracting partner Anthropic Ireland, Limited is located in the EU. However, Anthropic stores the data with subprocessors in the USA, among them the parent company Anthropic, PBC; the computation may also take place in other countries outside the EU. For this onward transfer Anthropic must provide appropriate safeguards (Art. 44 and 46 GDPR). According to its privacy policy, Anthropic bases transfers on adequacy decisions and standard contractual clauses. Information on this and a copy of the safeguards are available from us on request (section 1).
Paddle (United Kingdom): There is an adequacy decision of the European Commission for the United Kingdom (Art. 45 GDPR); the Commission renewed it on December 19, 2025. If Paddle transfers data to further countries, Paddle says it uses standard contractual clauses (see Paddle’s privacy notice).
Sending emails: Lettermint processes the emails in the EU; under the data processing agreement, a transfer to third countries is only permitted with our consent and in accordance with Chapter V GDPR.
8. Retention and deletion
We delete data as soon as we no longer need it for the purpose and there is no statutory obligation to keep it. The legal bases are stated with each processing activity in section 4.
The periods at a glance:
- Page view (4.1): not stored by the app; access log of the web server with shortened IP address 13 days.
- Error and operations logs (4.2): rotated by size, about 50 MB per service.
- Counters against abuse, including the check values of the contract forms (4.3): about 24 to 48 hours.
- Sign-in request with email address and check value of the link (4.4): until used, otherwise at most about 24 hours.
- Account (4.4): until the account is deleted.
- Agreement to the terms (4.4): as long as the account exists; with a purchase also in the order process (4.15).
- Session with IP address and browser identifier (4.5): until you sign out, otherwise 30 days after the last extension.
- Learning profile, time zone, appearance, language (4.6, 4.8): until the account is deleted; cookies see section 5.
- Learning progress, answers (including module check, “Practice” and the “Quick quiz”), learning days, instructions with all versions, evaluations (4.7, 4.9): until the account is deleted.
- AI background tasks (4.9): 2 days after completion.
- Check results with correction instruction and note; state of the phoenix per module (4.7, 4.10): until the account is deleted; check results of an own project together with the project.
- Own projects (4.10): until you delete the project or your account.
- Answers in the AI finder (4.12): not stored; only in the address, and therefore in the access log for 13 days.
- Trial lesson without an account on your device (5): on your device until you delete it, take it over or sign out; discarded after 30 days without a change.
- Quota counters, including for the monthly limits (4.13): until the account is deleted.
- AI cost rows (4.13): until the account is deleted, after that without reference to you and indefinitely.
- Shared AI cache, job suggestions (4.9): without reference to you, indefinitely.
- Data at Anthropic (4.9): at most 30 days; flagged content up to 2 years, classification up to 7 years.
- Emails at the sending service (4.14): 28 days, analysis events at most 90 days.
- Contact requests (4.18): until dealt with; statutory retention where applicable.
- Purchase data (4.15): until the account is finally deleted; accounting records until the end of the statutory period (accounting vouchers 8 years).
- Time the license was handed out (4.16): like the purchase data; the license key and offline edition are only with you.
- Order processes with agreement to the terms and request for immediate start (4.15): uncompleted ones 30 days; completed ones with the purchase, after account deletion without reference to the account until the end of the third calendar year after completion.
- Webhook messages from Paddle, shortened (4.15): 13 months; not tied to the account.
- Cancellation and withdrawal (4.17): until the end of the third calendar year after receipt, also after account deletion.
- Notices to us about declarations (4.17): in the app together with the declaration; summary emails in the mailbox as long as we need them for processing and follow-up questions.
- Reports and moderation (4.19): as long as we need them as proof toward authorities or in a dispute.
- Backups (4.20): on the server 7 days; externally encrypted up to about 3 months (7 daily, 4 weekly, 3 monthly snapshots).
- Deletion log with account identifier and times (4.21): 120 days after the final deletion or the reversal.
- Export file on request by email (4.21): until handed over to you, then deleted.
If you want to delete your account:
- How it works: in the settings under “Data and account” → “Delete account”, or by email to kontakt@kikoennen.de. You can export your data first (section 4.21).
- Immediately: Your account is blocked; you are signed out on all devices. We also immediately delete sign-in links you have not yet used. If you bought the course, your access ends; you get no money back because of this (within the withdrawal period you can withdraw first instead).
- After 30 days we finally delete the account: email address, language, country, time zone, learning profile, appearance, agreements, learning path, progress, answers, instructions with evaluations, check results, placement tests, badges, state of the phoenix, workshop with own projects, quotas, purchase data in the app and uncompleted order processes. If a cancellation or withdrawal has not yet been processed, we wait until that is done.
- What remains: the AI cost rows without reference to you; the webhook messages from Paddle, because they are not tied to the account; your cancellation and withdrawal declarations with the email address given in them and the completed order processes as proof until the end of the third calendar year, without a link to the account (4.15, 4.17); in the change log of the admin area, your account identifier for changes that concerned it (for example a reset quota); the deletion log with your account identifier for 120 days (4.21); reports and decisions on moderation. Your account identifier can remain in the error and operations logs until the entries are overwritten by rotation (4.2). Entries with their own period (counters, sign-in requests, background tasks) expire as described above. Your data disappears from the backups after about three months at the latest (4.20). At Anthropic and Paddle their own periods apply; Paddle stores payment data and invoices as an independent controller. Data we are legally required to keep is blocked until the end of the period.
- Reversal: Within the 30 days we reverse the deletion at your request (4.21). Your purchase is then back.
9. Do you have to give us data?
- Without an email address we cannot create an account.
- During setup the app only asks for your field (or “general”). Everything else is optional; without these details the app adapts its content to you less precisely.
- You can only use the AI features if we send the necessary texts to Anthropic (section 4.9).
- For a purchase, Paddle needs your payment details.
- For a withdrawal through “Withdraw from contract here” we need your name, the contract and an email address for the confirmation; for a cancellation through “Cancel contracts here”, your account’s email address. You can also withdraw or cancel by letter or email.
- You have no legal obligation to give us data.
10. Automated decisions
We do not make decisions based solely on automated processing that produce legal effects concerning you or similarly significantly affect you (Art. 22 GDPR).
These things happen automatically:
- Learning path: The app puts together your learning path by fixed rules from your work situation, goals and AI experience, without AI. It is a recommendation: you can start any module at any time; within a module the lessons open in order.
- Examples: The app selects pre-made examples and comparisons by your job and your field, without AI (section 4.9 c).
- Rule-based feedback and AI finder: The feedback on the Baukasten and the AI finder’s recommendation follow fixed rules, without AI. They are pointers for learning and for choosing a service and decide nothing about you.
- AI evaluation: The AI gives you feedback on your instruction. It feeds into your progress display. It is not a certificate or proof and does not decide on price, access or contract. You can revise your instruction and have it evaluated again.
- Limits: Quotas and protective limits apply automatically. They apply equally to everyone with the same status (free, course, AI extension) and do not assess you personally.
11. Your rights
Under the GDPR you have these rights:
- Access (Art. 15 GDPR): which data we process about you, and a copy of it.
- Rectification (Art. 16 GDPR): You change your learning profile, appearance and language yourself in the settings. You cannot change your email address there; write to us for that.
- Erasure (Art. 17 GDPR): see section 8.
- Restriction of processing (Art. 18 GDPR).
- Data portability (Art. 20 GDPR): the data you have given us, in a common, machine-readable format. In the settings under “Data and account” you download your data with “Export data” as a JSON file in a ZIP archive (section 4.21). On request we also send it to you.
- Objection (Art. 21 GDPR): see section 12.
- Withdrawal of consent (Art. 7(3) GDPR): Where processing is based on your consent, you can withdraw it at any time with effect for the future. What happened until then remains lawful.
- Complaint (Art. 77 GDPR): to a data protection supervisory authority, in particular in the EU country where you live or work or where the alleged violation took place. The authority responsible for us is the Landesbeauftragter für den Datenschutz und die Informationsfreiheit Baden-Württemberg (State Commissioner for Data Protection and Freedom of Information of Baden-Württemberg), Heilbronner Straße 35, 70191 Stuttgart, Germany (P.O. Box 10 29 32, 70025 Stuttgart), www.baden-wuerttemberg.datenschutz.de.
How to exercise your rights: Write to kontakt@kikoennen.de, ideally from your account’s email address. If we are not sure that the request comes from you, we ask you for confirmation. We reply within one month. In difficult cases the period can be extended by two months; we then tell you within the first month and give the reasons. As a rule this is free of charge for you.
You can fulfill access and data portability yourself with the export, and erasure with “Delete account” (section 4.21). Only Paddle has payment data; for this please contact Paddle (preferences.paddle.com or privacy@paddle.com).
12. Your right to object
Right to object under Art. 21 GDPR
If we process data on the basis of our legitimate interest (Art. 6(1)(f) GDPR), you can object at any time on grounds relating to your particular situation. This concerns in particular:
- the counters against abuse (section 4.3),
- IP address and browser identifier in sessions (section 4.5),
- error and operations logs (section 4.2),
- proof of agreements, order processes and declarations on cancellation and withdrawal (sections 4.4, 4.15 and 4.17),
- contact requests outside a contract (section 4.18),
- reports and moderation, insofar as they are based on our legitimate interest (section 4.19).
We will then no longer process the data unless we can demonstrate compelling legitimate grounds that override your interests, rights and freedoms, or the processing serves the establishment, exercise or defense of legal claims.
No particular form is required for the objection; the easiest way is an email to kontakt@kikoennen.de.
13. Data security
We protect your data with technical and organizational measures (Art. 32 GDPR), for example:
- encrypted transmission (HTTPS with HSTS);
- encrypted backups outside the server;
- separate database permissions: the web app may only read and write data, not change the structure of the database;
- no passwords; we store sign-in links only as a check value;
- cookies that scripts in the browser cannot read;
- strict security rules for content in the browser (Content Security Policy);
- example tools in an isolated area without internet access and without access to the app;
- access to the admin area only for people with the admin role;
- logs without personal data, as far as technically possible;
- limits against abuse (section 4.3), with check values instead of IP and email addresses for the contract forms;
- we make links in free texts of cancellations and withdrawals harmless in emails.
14. Minimum age
KI können is aimed at adults. Under our terms you may only use the app if you are at least 18 years old. The app does not check your age.
15. Changes to this policy
We adapt this policy when the app, our service providers or the legal situation change. The version published here applies. We inform you in advance of significant changes, for example by email.
16. Language versions
This policy is available in German and in English. The English version is a translation.
As of November 17, 2026