Ubenwa Health, a Canadian medical technology company, has launched Nanni AI, an AI-powered mobile app aimed at helping parents interpret their baby's cries to understand their needs better.
Parents can track their child's routines using Nanni AI, set feeding schedules, and log diaper changes through voice commands. The app also features a built-in chatbot for additional support.
Nanni AI has been trained on thousands of clinically labeled recordings of baby cries, translating over 1 million cries into English and French. The app boasts an impressive accuracy rate of 92.5% in detecting signs of birth asphyxia, a respiratory condition in infants.
Samantha Latremouille, Ubenwa co-founder, highlighted the overwhelming nature of initial doctor visits with a newborn, emphasizing the app's goal to empower parents by providing an easy-to-use tool to track and summarize their baby's health.
Founded in 2017, Ubenwa Health specializes in AI-powered sound-based medical diagnostics. The company has garnered support from notable figures like Yoshua Bengio, a Turing Award winner and deep learning pioneer, along with Radical Ventures and AIX Ventures.
The Montreal-based startup collected infant cry recordings globally to train the app. Initially obtained for research on infant neurological injuries, these recordings formed the basis for the Nanni app. Since its beta release in 2023, the app has been downloaded 100,000 times across 200 countries.
While currently available in limited languages, Ubenwa aims to make the app accessible in underserved languages, including Igbo native to Nigeria.
Charles Onu, Ubenwa’s CEO, views the sound of a baby's cry as a crucial vital sign. He sees Nanni AI as a foundational step towards making cry-based technologies widely accessible, aiming to integrate clinical-grade solutions for improved remote health monitoring.
Ubenwa is collaborating with hospitals to pilot its technology and is conducting more clinical studies to expand the app's capabilities to cover a broader range of medical conditions.
