Processing agreement
In short. The personal data on your website, and in what you send us for it, is yours. We only handle it to run your website and make the changes you ask for. We use AI only to carry out your changes, never to train AI models. We keep the data safe and, apart from the exceptions in section 7, in the EU, tell you quickly if something goes wrong, and hand it back or delete it when we stop working together.
- 1. Who and when
- 2. What data, and why
- 3. Only on your instructions
- 4. Confidentiality
- 5. Security
- 6. Other companies we use
- 7. Where the data is
- 8. Helping you
- 9. Data breaches
- 10. Checks
- 11. When the service ends
- 12. Liability
- 13. Duration and the rest
1. Who and when
This agreement is part of our terms and conditions and applies when we run a website for you on a subscription and, in doing so, handle personal data on your behalf. You accept it together with the terms and our privacy statement, by ticking the box in the onboarding form. You decide what is on your website and why: you are the controller. We handle it only for you: we are your processor (article 28 of the GDPR). Where this agreement and the terms differ about personal data, this agreement wins.
Some data is ours rather than yours: your account and your team's accounts in the panel, the sign-in codes, the record of each request and its AI cost (not the messages and photos in it), the record of who published which change, and payments and invoices. For those we are the controller, and our privacy statement applies.
2. What data, and why
| What we do with it | Store it, back it up, build it into your website and serve that to your visitors; read your change requests and photos and prepare the changes you ask for, with the help of AI (see below); keep earlier versions so you can restore them; and, when needed for support, look at it. |
|---|---|
| Why | Only to run the website subscription you have from us: hosting your website, making the changes you ask for, keeping it secure, and showing you its visitor statistics. |
| Whose data | Your website's visitors; people shown or named on your website, such as your team or clients in a review; and people shown in what you send us: in a change request, its photos or the onboarding form. |
| What data | Whatever you put on your website or send us for it — usually names, photos, job titles, phone numbers, e-mail addresses and reviews; what you fill in for your website in the onboarding form; your change requests (messages and photos) and what we prepare from them; and, in the access logs of the content delivery network, visitors' IP addresses and browsers, the pages they open and when. |
| For how long | As long as the service runs, within the periods below, and then as described in section 11. |
We put no cookies and no tracking scripts on your website: the visitor statistics we show you are counted from those access logs.
AI. We use AI only to carry out the changes you ask for, never to train AI models. Your request, its photos and the parts of your website it concerns are read by Mistral AI's models, through Mistral's EU service, to prepare the change, and automated rules check the result. AI can make mistakes, so you see every change in a preview first, and nothing goes live until you click Publish. Training is switched off in our account with Mistral. We have asked Mistral to keep nothing at all (zero data retention); unless it grants that, it may keep what we send and what its models return for up to 30 days, to detect abuse.
Only put people's names and photos on your website, and send them to us, when you are allowed to — for example because the people shown agreed. If what you send us includes special data, such as health details about a client, tell us first: it needs extra care, and you remain responsible for being allowed to use it. We don't use the personal data we handle for you for anything else: not for our audits, not to train AI models, and we never sell it.
How long we keep what
| What | How long |
|---|---|
| Messages in a change request | 90 days after the request closes |
| The original photos you send with a request | 30 days after the request closes (the versions used on your website stay on it) |
| What we prepared for a request: the AI's result, the list of changes and its working log | 90 days after the request closes |
| Previews | Until you publish or reject them, or they expire; at most 35 days |
| Published versions of your website | 100 days after they were published, so you can restore any version from the last 3 months; the live version for as long as it is live |
| Versions that were never published | 30 days |
| The version history of your website's files | As long as the service runs |
| The onboarding form | What you fill in moves into your customer record when you pay; the form itself is deleted 90 days after your website goes live |
| Raw access logs of the content delivery network | 35 days |
| Logs of the system that prepares your changes (only references to requests) | 30 days |
| Backups | 90 days |
3. Only on your instructions
We handle the data only as this agreement says and as you instruct us. Your instructions are the change requests you start in the panel, your click on Publish to put a change live, and what you ask us in writing (an e-mail is enough). Instructions from the people you let use the panel count as yours; you decide which of them may publish. If we think an instruction breaks the law, we tell you and may refuse it. If the law forces us to hand data to an authority, we tell you first, unless we are not allowed to.
4. Confidentiality
Everyone who can reach the data for us is bound to keep it confidential, and only gets the access the work needs.
5. Security
We protect the data with measures that fit the risk, including:
- encrypted connections (HTTPS) for every website, with certificates renewed automatically;
- your website stored with Amazon S3 in the EU and served through Amazon CloudFront, a content delivery network with built-in protection against attacks that try to overload a website;
- previews that are not public: they open only from the panel or through a personal link that expires;
- sign-in to the panel with a one-time code sent by e-mail, so there is no password to steal, and a notice e-mail when someone signs in from a new device; you decide who may publish;
- the AI can only change your website's own files, automated rules check its work, and nothing goes live until you click Publish; any published version from the last 3 months can be restored;
- our own systems (the panel, onboarding and our admin) hosted in a data centre in the EU, on a hardened Kubernetes platform (Talos Linux: a minimal operating system without remote login);
- every part of those systems reaches only the services it needs: network traffic is denied unless it is explicitly allowed;
- stored data and backups encrypted;
- passwords and keys kept encrypted, in a vault, never as plain text in code;
- malicious traffic to our systems blocked automatically (CrowdSec), and the systems monitored, with alerts;
- security updates applied regularly;
- daily database backups, with a copy kept off-site at another provider (see section 6);
- the management systems of our own servers (cluster, databases) reachable only by us, over a private network (VPN) with strong authentication — never from the open internet.
We may change these measures, as long as the protection stays at least as good.
6. Other companies we use
You allow us to use other companies (sub-processors) for the service. We bind them to the same data protection duties as this agreement. Today these are:
| Company | What for | Where |
|---|---|---|
| Amazon Web Services EMEA SARL (AWS) | Storing your website, serving it to your visitors, and its access logs; off-site backups; sending the e-mails we write ourselves from our mailbox, such as replies to your questions | EU; your public website is served from the nearest of over 750 locations worldwide (see section 7) |
| Mistral AI (Paris, France) | Reading change requests and photos, and preparing the change | The EU (Mistral's EU service) |
| Oracle Cloud | Our own systems: the panel, onboarding and our admin, where your requests and photos are kept, and the notice e-mails about your requests | EU |
| GitHub, Inc. | The version history of your website's files, and building the website from them | United States (see section 7) |
We tell you at least 30 days before we add or replace one. If you have a good reason to object, you may end the service from that date, and we refund the unused months.
Services from other companies that you ask us to add to your website, such as an embedded map or video, are your own choice: they are not our sub-processors, your arrangement with them applies, and your website's privacy statement should name them.
7. Where the data is
We store and process the data in data centres in the European Economic Area (EEA), with two exceptions that come with how the service works:
- Serving your website. Amazon CloudFront delivers your website to each visitor from the nearest of its more than 750 locations worldwide. Your website's public pages may therefore be cached, and visitors' requests (with their IP address) handled, at locations outside the EEA.
- Version history. GitHub, in the United States, keeps the version history of your website's files and builds the website from them.
Both happen under the safeguards the GDPR requires: the EU standard contractual clauses and, for the United States, the EU–US Data Privacy Framework. By this agreement you instruct us to use the sub-processors in section 6 in this way, including any access from outside the EEA by their staff under those same safeguards. Any other transfer outside the EEA happens only on your written instruction, with those safeguards.
8. Helping you
If someone asks us to see, correct or delete their data on your website or in what you sent us, we pass the request to you: you decide, and tell us what to do. We then carry it out, usually as a change to your website, and help where the service makes that possible. Data you have removed disappears from older published versions within 100 days and from our backups within 90 days.
If you ask, we also remove it from the version history, so earlier versions that contain it can no longer be restored.
We also give you the information you reasonably need for your own duties, such as a data protection impact assessment. For help that goes beyond the normal service we may charge reasonable costs; we tell you the cost before we start.
9. Data breaches
If personal data we handle for you is lost, leaked or reached by someone who shouldn't have it, we tell you without undue delay — we aim for within 24 hours of finding out — with what we know at that point. You then decide whether to report it to the Autoriteit Persoonsgegevens (within 72 hours) and to the people involved; we help you with that.
10. Checks
On request we give you the information you need to see that we keep to this agreement. If that is not enough, an independent expert bound to confidentiality may check it, at most once a year, with four weeks' notice and without disrupting our work. You pay for the check, unless it finds that we broke this agreement in a significant way.
11. When the service ends
This section also applies when we suspend your website for abuse or persistent unfair use, as the terms allow: we then refund the unused months, and you get your data so you can take it with you.
- Your website goes offline. We delete its published and unpublished versions and its previews. Your domain stays yours, so you can point it elsewhere.
- A copy for you. You can have a copy of your website's content, such as its texts and images, within 30 days after the service ends.
- Then we delete the rest: your website's stored files, including older stored versions, its version history at GitHub, your requests and their photos, and anything left of the onboarding form — unless the law requires us to keep it.
- Logs and backups run out by themselves, within the periods in section 2: the access logs within 35 days, anything Mistral may keep within 30 days, and our backups within 90 days.
Our own records, such as invoices and the record of each request and its AI cost, follow our privacy statement.
12. Liability
The liability limits in our terms and conditions also apply to this agreement. We remain responsible towards you for the work of our sub-processors as if it were our own. Each of us is responsible for its own duties under the GDPR; fines or claims that follow from your duties as controller are yours — for example when a photo of someone is on your website without their permission.
13. Duration and the rest
This agreement lasts as long as we handle personal data for you, and its duties on confidentiality and deletion continue after that. We may change it as described in the terms. Dutch law applies, and disputes go to the court named in the terms.