Privacy
This is a translation. The Dutch text is the official version.
DutchKlaar keeps almost nothing about you. Only to stop misuse, we keep a code of your internet address for one day. Here you can read what happens with your text and your exercises.
In short
- You do not need an account and you do not log in.
- DutchKlaar uses no cookies. We count how often the site is used, without knowing who you are.
- We do not keep your own text.
- Your progress and answers stay in your own browser.
What happens with your text?
You paste or dictate a text. We send this text one time to Claude. Claude is an AI model of the company Anthropic. Claude uses it to make the exercise.
Do you choose "Practise with existing questions"? Then your text only goes to our own server. The server uses it to find questions in our question bank that fit your text. Your text then does not go to Claude or any other AI. We do not keep it and do not write it in a log. This also does not count towards your daily limit.
Do you use "Speak your text"? Then your browser turns what you say into text. Some browsers, like Chrome, use an online service of the browser maker for this.
DutchKlaar does not keep your text and does not write it in a log. Anthropic does not use your text to train its models. Anthropic usually deletes the text and the questions within 30 days. Sometimes Anthropic keeps data longer: when the law asks for it, or when its safety system thinks there is misuse.
Anthropic may process data outside the EU, for example in the United States. For this, the EU standard contractual clauses in the data processing agreement of Anthropic apply.
Read more at Anthropic: commercial terms and data processing agreement.
Do not put your BSN number, bank details, medical details or other personal data in your text.
A photo, file or link
You can also take a photo, choose a file (PDF, Word or text) or paste a link to a web page. DutchKlaar takes the text out and puts it in the text box. You read the text and change it. Only when you press "Make my exercise" does the text go to Claude to make the exercise, just like a typed text.
We send a photo or a PDF once to Claude by Anthropic. Claude reads the text and sends back only that text. At Anthropic, the same rules apply to the photo or the PDF as to your text (see above). Your browser first makes a photo smaller. This also removes extra data of the photo, such as the place where you took it. We read a Word or text file ourselves on our server.
For a link, the DutchKlaar server gets the web page. The website then sees the address of our server, not yours.
DutchKlaar does not keep the photo, the file, the link or the text and does not write them in a log. Here too, your browser first solves a short puzzle, and the server remembers a coded version of your IP address, so nobody can do this very often. This code changes every day and is deleted after that day (see "Limits per day").
Do not take a photo of your passport, your bank card or other personal papers. Remove personal data from the text before you make the exercise.
What happens with the questions?
The AI is told to replace personal data in the exercise with made-up data. Then the AI checks every new question. Is the correct answer right? Does the question fit the level? Is there no personal data in it? Is there nothing hurtful? Only then do you see the question.
We keep the questions, not your text. They are in a shared collection and we use them again for other learners. The questions belong to DutchKlaar. Normally there is nothing about you in them, because the AI replaces personal data and checks every question. Do you still see something personal? Report it, and we will remove it.
We remove a question when learners report it or when we see a problem ourselves. The code of an exercise keeps working as long as the exercise is not removed.
Where does the sound come from?
The sentences you hear are made by ElevenLabs. ElevenLabs is a company that makes voices with AI. So it is a computer voice, not a real person.
We only send the sentences from the exercise to ElevenLabs, never your own text. We keep the sound and use it again for other learners. So each sentence goes to ElevenLabs only once. Does it not work? Then your browser reads the sentence aloud.
ElevenLabs may keep the sentences and the sound for some time. ElevenLabs processes the data in the United States. For this, the EU standard contractual clauses in the data processing agreement of ElevenLabs apply. Whether ElevenLabs may use the sentences to train its AI depends on a setting in our account. We turn this setting off.
Read more at ElevenLabs: privacy policy and data processing agreement.
Looking up a word
Do you tap a word? Then we send only that one Dutch word from the exercise to Claude by Anthropic, not your text and nothing about you. Claude gives the meaning in easy Dutch and in the languages of DutchKlaar. We keep that meaning and use it again for everyone. So each word goes to Claude only once.
The sound of a word is made by ElevenLabs, just like the sentences. We do not keep which words you look up. Your list "My words" is only in your browser.
Asking a question
At a question and after the exercise you can ask an AI something. Then we send your question, the earlier messages of that conversation and the questions of the exercise to Claude by Anthropic. Claude answers in the language you write in. If you write Dutch, Claude answers in simple Dutch. At Anthropic, the same rules apply to your question as to your text (see above).
DutchKlaar does not keep your questions or the answers and does not write them in a log. The conversation is only on your screen. When you close the page, it is gone. Here too, your browser first solves a short puzzle, and the server remembers a coded version of your IP address, so nobody can ask very many questions. This code changes every day and is deleted after that day (see "Limits per day").
Do not type personal data in your question, such as your name, address, BSN or phone number.
Checking writing and speaking
For a writing or speaking task, you can have your answer checked. We then send your answer once to Claude by Anthropic. Claude gives feedback in simple Dutch. At Anthropic, the same rules apply to your answer as to your text (see above). DutchKlaar does not keep your answer or the feedback and does not write them in a log.
For speaking, your browser turns what you say into text. Some browsers, such as Chrome, use an online service of the browser maker for this. We only get that text, never your voice. You can correct the text first.
Here too, your browser first solves a short puzzle, and the server remembers a coded version of your IP address, so nobody can do this very often. This code changes every day and is deleted after that day (see "Limits per day"). After that, your answer is only in your browser, with your progress.
Do not put real personal data in your answer. Use made-up names and addresses.
What do we keep for an exercise?
We keep the questions with their labels: topic, level, skill, words, type of question and estimated time. We also keep the title, the level, the date and the length of the exercise. This is stored with the sound in a database of Supabase, a company for data storage, on servers in the EU.
We keep nothing that shows who made the exercise: no text, no answers, no IP address and no information about your device.
Limits per day
Some things cost us money, because an AI does the work. That is why you can do them a few times each day: 5 new exercises, reading a photo, file or link 5 times, 10 answers checked, 10 questions with "Ask" and 30 new words looked up. Today's exercise, random questions and KNM from our collection have no such limit.
To count this, the server makes a code from your internet address (IP address). For this it uses a random key that is new every day. The database only holds this code, the date, what you did and how often. This code is a pseudonymised code of your internet address: with the key of that day, it can be linked to an IP address. That is why we keep it for at most one day. After the day we delete the codes and the key. Then nobody can link the codes to anything, not even we. We never keep your IP address itself and never write it in a log.
We only do this to stop misuse and to limit the costs. That is our legitimate interest (GDPR article 6(1)(f)). We use no cookies for this and keep nothing in your browser. So clearing your browser does not change the limit.
To recognise robots, we also look at a few signals: your browser solves a short puzzle, the page must really have been used (a tap, click or key), there must be a few seconds between opening the page and sending, and a hidden field must stay empty. If one network asks for very many puzzles, the puzzle becomes harder there for an hour. The server only keeps track of this in its working memory, with a coded network code. Of this we only keep numbers per day.
Continue on another device
You can make a link to continue an exercise on another device, for example on your phone. The link also has a QR code. Your answers so far are in the part of the link after the # sign. Your browser does not send that part to the DutchKlaar server. So we do not see it and we do not store it.
Anyone who has the link can see your answers. So send the link only to yourself.
When you report something
You can report a question or a whole exercise. Then we keep the type of report, the short reason you typed and which question or exercise it is about. We do not keep an IP address or a name.
To stop misuse, the server remembers a coded version of your internet address (IP address) when you report something or practise random questions or KNM from our collection. This way nobody can do this very often. This lasts at most one day and only in the working memory of the server. It is never saved and nobody can read it back. The same happens when counting, see below. For new exercises and the other things that cost us money, a day code is kept in the database, see "Limits per day".
For a new exercise, your browser first solves a short computer puzzle. This way we know without cookies that you are a person.
Do not put personal data in the reason.
Counting and error messages
DutchKlaar counts how often pages are opened and how often exercises are made and finished. These are only numbers per day. We do this without cookies and without anything that shows who you are.
To stop misuse, the server also remembers a coded version of your IP address here. This way nobody can send very many counts. This lasts at most one hour and only in the working memory of the server. It is never saved and nobody can read it back.
Does something go wrong in your browser? Then we send a short error message, without your text. First we remove everything between quotation marks, and also long numbers and email addresses. This helps us fix the error. We keep the numbers for 90 days and the error messages for 30 days.
We also use Vercel Web Analytics to count visits. Vercel is our hosting company. Vercel Web Analytics uses no cookies and does not keep your IP address. It gets the address of the page (without exercise codes, without search words and without anything after a ? or #), the page you came from, your country, and the type of device, browser and operating system. These are only numbers. Nobody can see who you are.
What is in your browser?
Your browser holds your progress and your answers for each exercise, and your choices under "Display" (text size, contrast, background, "Easy reading", sound and language). The level you choose (on the start page or after the level test) is there too. Your list "My words" is there too: the words you practise, with their box and date. The list of your recent exercises, for "Continue", is there too, and so is which lessons of your learning path you have finished. This is only on this device. The DutchKlaar server never gets it.
Do you put DutchKlaar on your home screen? Then the app does not keep any extra data about you for this.
Delete my data
With this button you delete what DutchKlaar has kept in this browser: your progress, answers, your list "My words", your level and your choices for display, "Easy reading", sound and language. Sound you have heard can still be in the normal cache of your browser. There is nothing about you in it: it is only sentences and words from the exercises.
The questions in the shared collection stay. They belong to DutchKlaar and normally there is nothing about you in them. Do you still see something personal? Report it, and we will remove it.
Your rights
Under the GDPR (in Dutch: AVG) you may ask which data we have about you and ask us to delete it. To stop misuse, we keep a code of your internet address for one day (see "Limits per day"). After the day we delete the code and the key, and then nothing of it is left. Apart from that, DutchKlaar keeps nothing that shows who you are. That is why we cannot find or delete any other data about you on the server. You delete the data in your browser yourself with the button above.
Questions?
Do you have a question or a request about privacy? Go to KlaarLabs. They make DutchKlaar.