Vision
AI that serves people and institutions without taking their data.
The point is not to make AI weaker. The point is to make it useful without turning every prompt, file, and draft into upstream visibility.
About
Keep your mind yours.
Tresor began with a simple discomfort: the moment an AI assistant can inspect your reasoning process, freedom of thought starts to erode. We are building the opposite.
The story behind Tresor.
It started somewhere ordinary: a co-working space at night. Two laptops open. One of us typed a half-formed idea into an AI assistant, just to sharpen it, then stopped mid-sentence.
That draft was not “content.” It was the inside of a mind: hunches, doubts, and the messy path to a better thought.
We realized how much private reasoning people had begun to share with AI assistants. If those tools can see inside the process, they can map preferences, learn blind spots, and quietly shape the next thought.
When thinking feels observed, people think smaller. That was the line we were not willing to cross.
Freedom of thought should not end at the edge of a textbox.
Tresor exists to give professionals and teams a workspace where powerful AI does not require surrendering drafts, knowledge, or internal judgment. Built in Luxembourg under EU law, it combines cloud convenience with a control posture that can actually be reviewed.
What we are building toward.
Tresor is not a protest against AI. It is an argument for AI that preserves privacy, proof, and operational simplicity at the same time.
Mission
A confidential workspace that is private, verifiable, and simple to run.
Tresor combines secure enclaves, customer-held keys, and reviewable controls with a product teams can actually adopt without running their own AI platform.
Operating posture
European legal footing with proof that survives procurement and security review.
Built in Luxembourg under EU law, Tresor is designed for organizations that need privacy commitments backed by runtime controls and evidence, not just policy copy.
Founding team
Different backgrounds, one shared conviction: powerful AI should become safer, simpler, and more accountable to the people using it.
CEO
Bob Krier
Bob drives Tresor’s business and partnerships. He’s spent his career helping ventures grow across Europe, blending strategic clarity with the pragmatism it takes to make things real.
Connect on LinkedIn
CTO/CPO
Dylan Thompson
Dylan leads engineering and product. Before Tresor, he built AI systems and created movies all over the world, combining technical expertise with creative vision to deliver impactful experiences.
Connect on LinkedInRecommended reads
Explore our latest writing on zero-access infrastructure, AI governance, and what it takes to keep digital autonomy intact.
6 min read
“We Don’t Use Your Data for Training” Is Not a Privacy Guarantee
That reassuring sentence covers only one narrow risk. Learn what it ignores, including retention, human access, profiling, breaches, subpoenas, and tool-connected data leaks, and what questions to ask instead.
Read the article
9 min read
Promises Aren’t Enough: Why Secure Cloud AI Still Isn’t Confidential
Many AI tools handling sensitive data rely on classic cloud security and contractual promises. Learn where this approach breaks down, how data can still leak, and why privacy-by-design architectures change the equation.
Read the article
8 min read
Your AI Chats Aren’t as Private as You Think
AI chat tools like ChatGPT and Claude can store and review your conversations. Learn how this works, the risks, and why privacy-first AI matters.
Read the article