FLORN analyzes structured data and separates what is established, refuted, and still unresolved. Same inputs, same result. Not a generative model. Public proof: historical FAERS extracts, 0 causal certification because the data do not allow it. https://florn.dev/results
Working with structured datasets, I kept seeing tools that produce confident-sounding answers the underlying data cannot support. I wanted the opposite: an engine that runs deterministically and states plainly what is established, what is refuted, and what remains unresolved, including returning nothing when the data do not allow a conclusion.