The Science Behind SpeakEQ

Coming soon

We’re putting together a detailed write-up on the research and models behind SpeakEQ: the acoustic features our system analyzes (pitch variability, vocal effort, speech rate, micro-prosody), the emotion and expression models that turn those features into feedback, and the published research our approach builds on.

This page will cover how we evaluate accuracy, the limits of what expressive-speech models can reliably tell you, and the design choices we’ve made to keep the feedback actionable rather than judgmental.

If you’re a researcher, clinician, or early user who wants to see any of this sooner, or collaborate on something adjacent, email us at support@speakeq.ai and we’ll share what we can.