Press X and the clone turns to glass. Every answer grows a living trust tree — source → consent → grounding → disclosure → synthesis, with confidence at each branch — beside a live model readout. JLens for a clone: nothing it says is a black box.
Concept mockup. The trust tree and model readout are illustrative of the interpretability layer — the same instrument idea behind jlens.arnao.ai, applied to a clone. Every node here would trace to real retrieval, consent state, and provenance in the live product. The point: a clone from the Responsible-AI lead shows its receipts, or it doesn't ship.