OQENYX Launches the Guardian AI Platform

After more than two years of development, OQENYX is going public with Guardian — the private, privacy-first AI platform we have been building from the ground up. Today marks the first public release, and we want to explain what it is, what it is not, and why we built it this way.

Guardian is our privacy-first AI platform — our own routing, authentication, inference, and product layer, built for developers and enterprises. Every architectural decision is grounded in one principle — privacy is structural, not cosmetic.

The first model in the Guardian family is Guardian 1.0 Thinking — a reasoning-optimized language model with a long context window, three selectable reasoning levels (low, medium, high), and minimal retention by default. Inference is encrypted in transit and at rest. There is no logging of prompt content. There is no training on your conversations. Guardian 1.0 Thinking is coming soon.

We chose to build Guardian with a privacy-first architecture because we believe the next decade of AI will be defined not just by benchmark scores but by the values that shape the platform. US hyperscalers are building towards lock-in. We are building towards independence.

Guardian 1.0 Thinking will power Onora, our conversational AI assistant. Get started at onora.app or contact us for enterprise onboarding.

This is day one of a much longer journey. We have a clear roadmap: Guardian Lite for low-latency workloads, Guardian Vision for multimodal use cases, and Guardian Edge for on-premises deployment in regulated industries. The foundation is solid. Now we build on top of it.

Thank you to everyone who supported us from the beginning — you made this possible.