An open-source rules-based framework for evaluating AI agent performance across various industries and use cases.

A minimal, code-first retrieval observability harness that measures why RAG systems fail to surface relevant evidence, without changing retrieval or generation.

Controlled experiment isolating reranking as a first-class RAG system boundary, measuring how evidence priority—not recall—changes retrieval outcomes.

Mini eval harness to turn “the model felt different” into measurable checks for honesty, calibration, prompt influence, and boundaries.