Sub-processors
Last updated 20 September 2026
When Cadre works for a client, the client is the controller of their customers' personal data and Cadre is the processor. These are the companies Cadre uses to do that work, what each one does, and where each one processes. Article 28 of the GDPR requires a client to know this before they authorise it, so it is published rather than sent on request. Some entries below serve Cadre's own website rather than a client's data. They are listed anyway, so that the list is complete.
Stored in the European Union
These three hold the records the service creates.
- Cloudflare. Hosts the site, stores the records the service creates (chat conversations, enquiries and intake records), and runs the challenge that keeps spam out of the forms. The storage is pinned to the EU, and the code that answers a request runs on Cloudflare's network at the location nearest the visitor.
- Resend. Sends the enquiry and lead email that reaches a client's inbox. The account is set to Resend's EU region.
- Plausible. Counts visits. It is cookieless, sets nothing on a visitor's device, builds no profile, and is hosted in the EU.
Processing Outside the European Union
Three companies process personal data outside the EU. Chapter V of the GDPR permits this where the transfer rests on an approved safeguard. Anthropic, ElevenLabs and Google are each certified under the EU-US Data Privacy Framework. Anthropic and ElevenLabs additionally incorporate a data processing agreement with the standard contractual clauses into their terms.
The next section sets out what is kept away from the model, and where that limit ends.
- Anthropic. The model that writes the replies in the chat, builds an instant demo from a public website, and writes the summary of a conversation afterwards.
- ElevenLabs. The voice on the phone line and on the instant demo call.
- Google. Reads a calendar's free and busy times and writes the appointment when a visitor books, and completes an address when a visitor types one.
What Is Kept Away from the Model
Where Cadre runs a campaign for a client, the names and contact details are merged in on Cadre's own systems after the text is written. The model writes the message; it is not given the list of people the message goes to.
Anthropic states that data sent through its API is not used to train its models. Cadre does not use a client's data to train a model of its own, and each client's data is kept separate from every other client's.
The chat is where that limit ends, and it is worth stating plainly: the chat box is free text, so anything a visitor types into it does reach the model, including contact details typed there instead of into the form that opens for them. What goes into that form posts straight to Cadre, and the model never sees it.
Changes to This List
Cadre tells its clients before adding or replacing a sub-processor, so a client has the chance to object before the change takes effect. This page carries the date it was last changed.
Questions about anything here go to hello@cadre.lu.