What is ChatterBaby?
ChatterBaby compares your baby’s sounds to our database of roughly 1,500 sounds, to determine whether and why your infant cries. It uses math(emagics). Before you use it, you must apologize to every math teacher you’ve ever told, “Why do I need to learn this junk?”

How accurate is ChatterBaby?
ChatterBaby correctly identifies roughly 85% of pain cries, and is roughly 90% accurate for catching any baby cries. Fussy/whiny cries are the cries most often mistaken for non-crying.

How can I make ChatterBaby work better?
The less background noise is present, the better the algorithm will work. If you feed it a soundclip of you singing to your crying baby, it will not work. If you feed ChatterBaby a soundclip of your dog barking, it will not work.

What kinds of cries does ChatterBaby predict?
The algorithm assumes the baby is crying for only three reasons: hungry, fussy, and pain. If your baby is crying because he really, really, wants to taste some Legos and you just won’t let him (true story), it will not work. Sad fact: separation anxiety cries are predicted as “pain” by our algorithm.

Can I trust the ChatterBaby algorithm?
Your brain and your own intuition are far more powerful than the fanciest equations. If the ChatterBaby algorithm and your own common sense disagree, always trust your brain.

What happens to my data?
We store it for science on a server that is HIPAA-compliant, removing as much information as possible that links your data to you individually. We are interested in discovering whether abnormal vocalization patterns in infants can predict neurodevelopmental delays such as autism. Odds are, a human being will never actually hear your baby’s audio sample; a computer script will run some math on it and throw the answers in a big spreadsheet to do even more math. For more information, see the Consent Form you agreed to when launching the app: link to consent.

Is there a warranty?
Nope. Nein. Zero. Zilch. Nada.

Is this a medical device?
No. See also “Can I trust the ChatterBaby algorithm?“ above.

What about remote monitoring?
TBD.

UCLA Health

Latest release