The Lab

oneg.ai is an applied AI research lab. We produce frameworks, ontologies, datasets, and papers grounded in multiagent systems, belief revision, and knowledge representation.

The name comes from お願い (onegai), a Japanese noun meaning "request" or "favor," commonly used as an interjection equivalent to "please." For us, it maps to a posture: the polite but firm insistence that intelligence be grounded in rigor. Our tagline, Intelligence, please, is both a literal reading and a statement of intent.

That intent shapes everything we build. We treat agent reasoning as a problem of logic, not prompt engineering. We treat knowledge as something with structure that can be validated and explained, not just retrieved. And we treat reproducibility as non-negotiable: every experiment runs on consumer hardware, every framework ships as open-source software.

The lab operates as a small, focused research group. We bring in domain specialists and engineers as projects demand, keeping the core tight and the expertise deep. Some of our work feeds directly into products built by partner studios; most of it ships publicly.

If any of this resonates, we'd like to hear from you.