Data stays home
Our products assume from the first line of code that your information never needs to leave your hardware.
Company
Local-first, by design.
JMNI Labs is an independent company building private AI products around a single conviction: capable AI belongs on hardware its users control.
The models caught up. The hardware caught up. What's missing is software that lets people use private AI where it matters most — in the law office and the classroom.
Principles
How we work.
Measured, not claimed
We run everything on our own lab first. If we haven't measured it, we don't say it.
Grounded answers
AI that cites its sources and admits its limits is worth more than AI that sounds confident.
Publish the work
Findings, checkpoints, and field notes go out in public — including what didn't work.
Team
Small team, serious lab.
J
John Gutierrez
Founder
An engineer with years of experience building private inference systems, John founded JMNI Labs to turn that work into applications people can use, and leads research and product.
JMNI Labs is built with a close circle of developers and collaborators from the open-model community. We're incorporated in Hawaiʻi and work remote-first.
Roadmap
Where we're headed.
Now
Design partners
- Ward with early law-firm design partners
- Recitation private preview
- Ongoing research on local serving
Next
First firms
- Ward pilots with litigation and PI firms
- Redundant, colocated inference
- Recitation v0.1 public release
Later
Scale
- Ward local mode for firms with their own hardware
- Independent security audit
- More practice areas on the same foundation
Let's talk.
Law firms, schools, learners, researchers — we'd like to hear what you're working on.