Last released May 24, 2026
Deterministic semantic assertions for LLM tests. Freeze a judge's verdict once; replay it forever.
Supported by