Axom Labs · MMXXVI

Intelligence you can grow.

A new architecture for machine intelligence. Built in America, running on your hardware, and yours to own.

The Axom app

Your AI. On your machine.

Chat, projects, background jobs, memory, and personality, all served by Axom's own runtime on your hardware. Prefer the terminal? The same session lives there too.

The Axom app: a chat ready for a new conversation, with Projects, Jobs, Network, Memory, and Personality in the sidebar, running X1-Ultra locally.

Memory you can read

See what Axom believes about you. Every note cites where it came from, and nothing sticks until you accept it.

Jobs that keep going

Hand off a goal or have it keep watch. Jobs run in the background and survive a restart.

A voice you set

Tune how Axom talks to you (tone, length, humor, format) with a live preview reply.

X1-Ultra, measured
119/ 120
Test scenarios across research, files, and code, carried through to an answer.
~4B
Parameters active per word, out of 26B.
63.9tok/s
Generation on a desktop NVIDIA DGX Spark.
5tiers
From 5 GB of GPU memory to 20. One architecture on all of them.
American made · Built for the people

Intelligence that belongs to the people who use it.