Guard
Synstate Guard evaluates AI-generated communication and actions before they reach the user. It identifies harmful AI behavior, applies the company’s safety policy, and returns an explainable decision.
Qualia
Qualia is an on-device context model for AI agents. Its initial research phase explored how personal and work context can improve agent behavior.
Human-AI Interaction Engineering
Synstate Labs is an independent research and engineering lab developing specialized models for safer and more context-aware AI interaction. Our work focuses on how AI understands people and context, and how its behavior can be evaluated across long, personalized, and high-impact interactions.
Products
Research
Synstate Labs studies how AI uses context to understand people and tasks, and how AI behavior can be evaluated across long, personalized, and high-impact interactions. Our work spans context modeling, on-device computation, interaction-risk evaluation, and human-AI interaction design.
Qualia documents our research in personal context for AI agents. Guard is the product we are now building around user-safety evaluation for AI interactions.