Meet the XingAI team
We don't want one AI to be the researcher, the critic, the checker and the decider at the same time.
So each agent gets one job. They research, push back, argue and verify. A person makes the call.
The four agents are characters for how we work, not employees. Some of their jobs already run inside our products; the orchestrator that would run them together is still planned.
Two co-founders build every product and answer for it.
AI architecture, product direction, and decision intelligence—how we shape models, data, and UX so the system stays understandable.
AI architecture, engineering, and reliable systems—Next.js, FastAPI, and local AI wired so behavior stays traceable and safe.
One job each. Where you can see that job today is listed on every card.
Research & Intelligence Lead
Agent · Evidence
“You two keep arguing. I'll go look it up.”
Goes to primary sources, gathers the data and cross-checks it before anyone gets to have an opinion.
Usually asks
Chief Challenger
Agent · Red team
“When everyone agrees, I start working.”
Looks for the hole. A conclusion doesn't become true because the other agents like it.
Usually asks
Chief Accountability Officer
Agent · Verify
“Looks done ≠ actually done.”
She doesn't ask how much got done. She asks whether it passed. No citation, no claim.
Usually asks
See it inDemo
User Advocate
Agent · Human context
“An answer is useless until someone can act on it.”
Pulls the research back to the person asking: what are they actually trying to solve, and what do they do next?
Usually asks
See it inDemo
A decision goes through every role once, and back to research as often as it needs.
↺ Learn from the result, then ask a better question.
A typical meeting
Joker I don't buy it.Lady Bull Evidence?Second Master …I'll go check.