Kintsugi shuts down and opens its depression detection AI work

In short: Kintsugi says it is shutting down its commercial operations after it could not get FDA clearance in time, and it will release its AI voice models and research to the public.

What happened

Kintsugi is a startup based in Berkeley, California. It was founded in 2019 and has spent years building software that looks for signs of depression and anxiety by analyzing a short speech sample.

On February 11, 2026, the company announced it is shutting down its commercial business. The CEO and founder, Grace Chang, said the company could not secure timely FDA De Novo clearance, which is a pathway the US Food and Drug Administration uses to review new kinds of medical devices.

Kintsugi raised about $30 million from investors, including Insight Partners and Acrew Capital. Now it says it will place its AI models, scientific methods, early research, and published studies into the public domain, meaning anyone can use them without a paywall.

Why it matters

Mental health checks today often rely on questionnaires and conversations. Kintsugi was trying to add a more objective signal by using “voice biomarkers,” meaning patterns in someone’s speech that might relate to health (like noticing a cough can hint at a cold, but using many tiny voice clues instead). The shutdown shows how hard it can be to turn health AI into a real product, because regulators and the market can move slower than startups can afford. By releasing its work publicly, Kintsugi may still influence future tools, even if they are built by other teams.

Source: The Verge AI