About ROSA

ROSA is a research project focused on building safe, autonomous AI agents. We are developing a recursive operating system that can learn, adapt, and use tools independently — all while keeping sensitive data local and secure.

From continuous learning and memory management to complex tool use like CAD modeling, ROSA pushes the boundaries of what agentic systems can achieve. Our goal is a future where AI assistants handle workflows autonomously and safely.

ROSA System UI

This demo shows how a local AI system evolves from simple chat into a reliable, tool-using agent. It introduces the key idea behind agentic AI: not just generating text, but completing tasks in a controlled and secure way.

You'll see local execution, tool use, browsing, file access, memory-based authorization, CAD examples, and a trainer agent that extracts valuable information from chats and documents to improve the system over time.

Get In Touch

Building the Next Generation of Secure AI — Together

ROSA is currently in an advanced development stage and is being prepared for real-world implementation.

We are looking for a pilot partner in the Netherlands to test and refine the system in a practical business environment. The goal of this pilot is to validate workflows, measure impact, and further develop the platform based on real operational use.

After the pilot phase, we will expand and scale the platform based on the lessons learned.

If your company is interested in secure AI automation and wants to help shape the future of controlled AI systems, we would be happy to explore a pilot collaboration.