At bl0x, security and privacy aren't a sales pitch — they're a design principle. We'd rather ask the hard questions during design than after an incident.
The AI demos on bl0x.be — image recognition, speech-to-text — run entirely in your browser, using your own processor or GPU. No photo, voice or text is sent to our servers or to third parties. What stays on your screen, stays there.
This is no accident — it's a deliberate choice that illustrates how we think. AI doesn't have to mean your data disappears into the cloud. In many cases, intelligence can run closer to the source: on a server you control, or even on the device itself.
Security isn't a feature we bolt on afterwards. At every layer we think about access control, data flows and what could go wrong — and what we do about it.
We work on the principle of least privilege: a system only gets access to what it truly needs. No broad API keys, no unnecessary copies of sensitive data.
Processing personal data always has a clear legal basis. We help clients think through purpose limitation, retention periods and the right to erasure — even when those questions are uncomfortable.
We don't build black boxes. If an AI system supports a decision, a human must understand why. We document the logic, limits and edge cases — for your team too.
After handover, the client manages the system independently. No hidden dependency on bl0x infrastructure, no data left lingering with us. You own what we build together.
Every system we hand over comes with clear documentation: what it does, what it doesn't do, how to switch it off if something goes wrong. Not a manual for engineers — one for the people who use it daily.
We don't promise perfection — it doesn't exist. What we do promise: we name risks honestly, don't take shortcuts to hit a deadline, and never push clients further than what's safe and responsible. If something can't be done responsibly, we'll say so.
Discuss your situation →