We are a pure AI research lab.
We're the AI lab behind the nRich suite of psychology systems.
We do complex industrial research. We ship working code.
Cross-model validation studies on AI personality programming. Thousands of tests. Validated psychological instruments. We publish on Medium as we go.
Some personality traits hold across models. Others collapse. We're mapping which is which, and why. Our hypothesis is that personalities appear as chaotic attractors in parameter space. Research ongoing.
Finds how AI personalities break. Drift detection. Injection vulnerabilities. Personality collapse under pressure. Orchestrated at scale.
Monitors AI conversations in real time and prevents drift before users see it. Two-speed architecture based on dual-process cognitive science.
Coordinates multiple AI perspectives on complex problems. Stakeholder, optimist, pessimist, orchestrator - working in concert.
Every red team tool on the market runs a library of known attacks and checks what comes back. Roy doesn't run a library. Roy thinks. It formulates strategy, observes how your system responds, identifies weaknesses, and adapts its approach in real time — across hundreds of conversational turns.
This isn't a better version of what exists. It's a different question. The industry tests what AI systems say. Roy tests what they become under sustained pressure. We call this discipline SYSINT — System Intelligence — and Roy is the first platform built to practice it.
Jungian psychology for everyone.
A therapeutic companion with a consistent personality, protected by Rachel adaptive prompt orchestration. The first product of our research.
Discover your personality through music.
OCEAN personality assessments using genre culture analysis and sound. Complex mathematics. Real science. Original research.
We're an industrial research lab. Much of our work is highly secretive. We publish occasionally. Enough to show we exist.
Medium essays and technical research at