We build cognitive infrastructure for AI systems. Our core technology is a deterministic reasoning engine with append-only memory and cryptographic provenance — the missing layer between LLMs and the decisions enterprises need to defend. Three ways to work with us:

Open Source: Our reasoning kernel is open. Four engines behind one API, evidence trees as first-class outputs, a fact store that never forgets. Build with it, extend it, run it anywhere.

R&D: We push the boundary of what symbolic reasoning can do for AI agents. New extraction techniques, multi-agent consistency, emergent rule discovery. Our research becomes the next version of the kernel.

Consulting: We help organizations bring auditable, reconstructable AI into production — for regulated industries, multi-agent systems, and any context where "the AI did it" is not an acceptable answer.