About
We measure what everyone else asserts.
Tesseract Semantics builds and validates the ontologies AI systems reason over. Every claim we make is backed by a study we published, with the code that produced it.
Why this exists
Every vendor in this category claims their system is more accurate. Almost none of them publish the method, the data or the code, so a buyer has no way to tell a real improvement from a confident sentence.
We work the other way round. We point our own tools at real public ontologies, registers and standards, and we publish what we find. That includes the findings that went against us, the experiments that failed, and the checks that turned out not to work. A body of evidence someone can check is worth more than a claim they cannot.
How we work
- We treat the ontology as infrastructure, so it is versioned, validated in your pipeline and reproducible from source rather than living in a modelling tool and a person head.
- We build in the open. The engine is MIT licensed and the studies are public, which means you can audit the thing you are buying before you buy it.
- We hand back capability rather than dependency. If we disappear, the model, the tests and the evidence trail still work.
Who we are
A small technical team working on ontology engineering, standards crosswalks and AI assurance, with delivered work across government, research and industry. The engine and the studies are the useful part, so they are what we put in front of you.