Every Truverai evaluation reports eight metrics in a fixed order, made up of 32 sub-metrics: Truthfulness, Relevance, Usefulness, Verifiability, Evidence, Responsibility, Authenticity, Integrity.
Truverai scores are reproducible within a published lock: engine v4.7, prompt hash 1a6c25c7e25ddbf8, model google/gemini-3-flash-preview at temperature 0, and a median of three runs. That is reproducibility within a fixed configuration, not a claim that a different language model would return the same number.
Parity controls keep evaluations comparable. The prompt is hash-locked, the model and temperature are fixed, each evaluation is a median of three runs, and a run group whose scores disagree beyond the allowed tolerance is voided rather than published.
Truverai evaluates how a communication is presented, not whether a belief is objectively true, and it scores what was said rather than who said it. Plan and payment tier do not affect scoring: anonymous, free, and paid evaluations run the identical path.
This page needs JavaScript for the full interactive view. The summary above is the same information in plain HTML.