What we’re building
The family of systems.
Everything here is real and in motion — and labeled with exactly where it stands. Nothing on this page is further along than we say it is.
Staff Assistant
Your team’s questions, answered from your own playbook.
A private assistant for clinical and front-office teams. It answers only from documents you’ve approved — protocols, policies, onboarding guides — with a citation back to the exact source and a named escalation path when a question needs a human. Nothing is pulled from the open internet, and your documents never train anyone else’s model.
Where it stands: taking demos now. Bring one real protocol and watch it answered.
Book a 15-minute demoJRL Stories
Family stories, kept in the voices that told them.
A voice-first app for family storytelling. Someone asks a question, someone answers it out loud, and the story is kept — playable within the family, in the teller’s own voice. The AI arranges recordings and smooths the seams between them; it never invents a word of what a person actually said.
Where it stands: in family testing now. Not yet open to the public.
Aether
Private intelligence for family wealth.
A private intelligence system for families whose wealth has real complexity — entities, advisors, documents, decades of history. It aggregates, researches, and surfaces what’s true, with a full audit trail, and it is designed to run inside the family’s own controlled environment. It never advises, never recommends, and never touches money — every decision stays with the humans it serves.
Where it stands: in private development.
VersaStage
Real lyrics. Real musicians. Real stories.
A music-making marketplace and music player, in the lab. Songwriters put up lyrics, artists answer them with real recordings, and the people listening decide what rises.
Where it stands: in the lab. The frame is built and the catalog is honestly empty — the first song isn’t written yet.
A public-records transparency engine — sourced, neutral facts about public institutions, every claim linked to a verifiable public record. Early research; everything in it today is synthetic sample data.
Follow the build.
Occasional notes as these systems grow — written by a human, sent manually, never spam.