Honest speech analysis from your real voice

As you rehearse, Rhetoric measures the things you can't feel: your pace in words per minute (including where it drifts across the speech), how often filler words slip in, how much of your material you actually covered, and how quickly you recovered when you lost your place.

These are computed locally from your speech (fast, private, and consistent), so the number moves only when your delivery actually changes.

Beyond delivery, the Rhetorical Audit analyzes the writing itself (ethos, pathos, and logos; structure and narrative; clarity and diction) and quotes the exact phrases it's reacting to, with concrete rewrite suggestions.

The overall persuasion score is computed in code from per-dimension ratings, not improvised by the model, so it stays stable and explainable: you can see that, say, logical rigor is what's holding the score down, and exactly which sentence to fix.

This is the rule the whole product is built on. Numbers measured locally from your voice are shown as measured. Judgments made by AI are shown as AI judgments. And when no speech is captured, Rhetoric tells you that instead of inventing a score. Trust is the point of measuring at all.

Analysis isn't the end of the loop; it's the hinge. Specific coaching notes turn each session into a sharper draft, which you rehearse again, measure again, and watch improve. Prepare, perform, measure, improve: the report exists to make the next run better.