Hi!, I'm Art, a scientist who spent 25 years presenting research and still watched the perfect answer show up five minutes after the meeting ended. The pattern that finally made me build this: I answered closed questions with open lectures. "Will it be ready by Friday?" is a yes-or-no question, and I gave it three paragraphs of methodology. Every time. Knowing the material was never the problem. Delivery is a motor skill, and I had spent zero minutes training it. So Dictalus is a gym, not a course. You get a prompt, record a short take, and get numbers: words per minute, fillers per minute, deliberate pauses, pitch variation in semitones (so "you sound monotone" is finally falsifiable), and a segment-by-segment arc of your certainty and tone; where you committed and where you slid into "sort of, I think, maybe." Then one fix. Then you run it back. Things I cared about that you might too: — Privacy is the design, not a setting: recordings are deleted the moment scoring finishes by default, and the live meeting meter runs entirely on-device. Nothing leaves your phone, unless you want to do more with it. — The analyzer and live meter are free without an account. Feed the analyzer the Gettysburg Address: zero hedges in 272 words. then feed it your last project update. That comparison is the whole product in ten seconds. — Free is real: 5 scored reps a day forever, full curriculum included. You can subscribe to get a more powerful model and faster results. Happy to answer anything about the speech metrics (under the hood) or the build.

Dictalus

A private rep gym for speaking under pressure

A private rep gym for speaking under pressure