voice-biomarkers
(2)
Kintsugi
Kintsugi is a Berkeley-based AI company that built voice biomarker technology able to flag signs of clinical depression and anxiety from roughly 20 seconds of free-form speech. Founded in 2019 by Grace Chang and Rima Seiilova-Olson, it sold an API-first platform into call centers, telehealth, and remote patient monitoring before winding down operations in early 2026 and releasing its models, methods, and research into the public domain.
Grace Chang
Grace Chang is the Taiwanese-American founder and CEO of Kintsugi, the Berkeley-based startup that built AI capable of detecting clinical depression and anxiety from just 20 seconds of free-form speech. A five-time entrepreneur with roots in signal processing and consumer tech, she raised $28 million, launched Japan's default mental health screener, and built the world's largest annotated voice dataset for mental health machine learning - before making all of Kintsugi's research and technology open source in early 2026 after FDA regulatory hurdles made the venture-backed model unsustainable.