Last updated: 22 August 2026
This page is written for the person at your school who has to sign something off: a head, a DPO, an IT lead. It says what the product does with data in specific terms, so you can check it rather than take our word for it. The privacy policy is the formal document; this is the plain account behind it.
KnowThinkLearn is a concept-based curriculum tool built on a coaching practice. AI was used to help write the software, in the way a compiler or a code library is used. No part of the running product sends anything to an AI service. Not your unit plans, not your teachers' judgements, not a word about any student.
That means there is nothing to opt out of, no model to configure, and no setting a teacher could switch on by accident. It also means no AI runs in the background: nothing fires on a schedule, on save, or when a page loads. Every automated job in the system is a nightly check that emails us if the software threw errors, and it reads error messages and page paths only.
You can verify this yourself in about a minute. Open the browser's network tab, use the planner normally, and look at where the requests go. Every one lands on our own domain or our database host. There is no request to any AI provider, because the code no longer makes one.
The strongest privacy protection is not encrypting a student's name well. It is not having the name at all.
So our preferred practice, and what we ask schools to adopt in assessment and moderation, is this: identify students by a number that is unique within the class or the year group, not by their full name. Keep a separate list mapping numbers to pupils, held wherever your school already holds its roll, and never in here.
Two details matter. First, use a class or year-group number, not the school-wide student ID, because a school ID follows a child between systems and is genuinely identifying, while "17" in a class of thirty is not. Second, keep comments free of identifying detail too: a name in a free-text note reaches the same place as a name in a name field. The workspace flags likely slips as you type, including anything shaped like a full name or a long ID number, but the flags are advice, not a lock.
Where a first name is genuinely needed to make a report readable, the assessment workspace asks for a first name only and says so.
Our database is hosted with Supabase in Paris, France (eu-west-3). The websites are served by Cloudflare.
The great majority of what we hold is curriculum, not people: unit plans, statements of inquiry, criteria, learning experiences. Personal information about students appears only where a teacher types it, and it appears in more than one place, so it is worth being precise:
That last point is the reason we would rather you never typed a name than trusted us to tidy one up afterwards.
Access is enforced by the database itself, on every query, not by the screens in front of it. For any record about a student, that means two people: the teacher who created it, and a coordinator at the same school. Nobody at another school can reach it, and access also depends on your school's subscription being live.
We do not read your content to improve the product, and we do not use it for advertising or sell it. We can see it only in the ordinary sense that we administer the database, for example when you ask us to fix something.
The pages that display learner progress and moderation records carry two protections beyond ordinary sign-in, aimed at the unattended laptop in a staff room:
Both sit on top of the database rules described above, which remain the real boundary.
There is no third-party analytics or advertising code anywhere on this site. No Google Analytics, no Tag Manager, no advertising pixel, no session recording.
We count page views ourselves, and those counts record the page, the referrer and the browser type, with no account identifier attached, so they cannot be traced back to a person. Error reports record what broke and on which page, and carry no account identifier and none of your content.
For completeness, two external files do load in the browser: the payment provider's checkout script on the checkout page, and web fonts from Google, which means Google can see the visitor's IP address when a page loads. Everything else is served from our own domains.
None of these receives student records in the ordinary course of the product working. There is no AI provider on this list, and that is the point of section 1.
We would rather be plain about this than reassuring. Our database is in France. If your school is elsewhere, for example in Korea, Japan or the Gulf, then using the product means personal data is transferred out of your country and stored in the EU. Under laws such as Korea's PIPA, that transfer normally needs to be disclosed to the people concerned and to rest on a proper basis, and information about children carries additional requirements, including guardian consent below a set age.
Two practical consequences. First, the class-number practice in section 2 does most of the work here, because a number that means nothing outside your own roll is a far weaker piece of personal data to transfer than a child's name. Second, this is a conversation to have with us rather than around us. If your school has a data protection questionnaire or a processing agreement to review, get in touch and we will take it from there.
You can delete a plan or a record in the product at any time, and deleting a plan removes its version history with it. If you want your whole account and its content removed, email us and we will do it within 30 days, apart from transaction records our payment provider must keep for tax and accounting.
Anything on this page, and anything it does not cover: contact@knowthinklearn.com. If you are evaluating us for a school and need something in a particular format, say so and we will write it that way.