Independent
by design.

A clear speciality. We focus on compact models, document knowledge and evidence. We judge progress by how well a model performs those tasks.

Claims that can be checked. We separate early experiments from validated capabilities. Comparisons need a stated setup, matched conditions and visible limitations.

Efficiency as a design constraint. We consider compute, memory and deployment alongside accuracy. A useful model should fit the conditions in which it will be used.

We welcome conversations with researchers, collaborators and teams working with document-intensive problems.