Build
Products people can trust.
OOMU, our first mass-market product, is a local-first AI workstation that runs everyday work on your own Mac and shows you the plan before it acts.
Eldris · An applied AI lab
We build AI products people can trust, help organizations deploy AI with confidence, and turn applied research into working software.
Where we come from
Eldris began by securing AI systems for organizations that couldn't afford to get it wrong. That discipline never left. Every product we build, every engagement we take on, starts with trust and safety.
What we do
Build
OOMU, our first mass-market product, is a local-first AI workstation that runs everyday work on your own Mac and shows you the plan before it acts.
Advise
We help organizations decide where AI belongs, put it into production responsibly, and govern it once it's there. The advice comes from work we've shipped ourselves.
Research
We experiment with local-first agents, hybrid model economics, and AI security. The experiments that work become products. OOMU started this way.
Our first mass product
Trust, built in.
OOMU began as a tool we built for our own work inside the lab. It runs everyday tasks on your own Mac, reaches for the cloud only when a task truly needs it, and asks before it acts on anything that matters.
It's free, open source, and simple enough for people who have real work to do. An Enterprise edition adds the deployment and controls teams need.
For organizations
They fail on unclear strategy, ungoverned rollouts, and tools that never fit how people actually work. We've seen it from every side: as builders, as operators, and as the people called in afterward.
Eldris advises organizations on where AI helps, how to deploy it without creating new risk, and how to govern it so it stays worthy of trust.