What is Agubu?

The short answer to what is Agubu: a baby cry analyzer, care logging and soothing sounds app for parents, on iPhone and Android, in English and Turkish. The long answer is why it differs from the rest of the category: Agubu accepts that the reason for a cry cannot be read from its sound, and builds the product on that fact.

The name comes from the Turkish root agu — the sound infants make when they are content. It reads the same in Turkish and English.

Who makes it

Agubu is made by Yiğit Samet Ölmez, an independent developer living in Turkey. There is no company, no investor and no team behind it; the person answering support emails is the same one. We say this not as a shortcoming but as the explanation for why the product is so plain: an app that runs no server, opens no accounts and sells no ads is one a single person can keep going.

Why it was built

The baby cry analysis category is known for two things: high promises and low store ratings. The complaints repeat in the reviews: “it gives a result even in a silent room”, “it gave four different answers to the same cry”. We did not just read those complaints; we measured them.

On the only public dataset labelled by cause, with baby-wise cross-validation, predicting the cause from sound stayed at chance level, while the same pipeline recognised the individual baby far above chance. A peer-reviewed 2023 study reaches the same conclusion. The model inside one app on the market named a cause for absolute silence with 100% confidence. We built the product on that measurement: an app that does not pretend to read the reason from the sound. All the numbers are on the Methodology page.

How it decides

The microphone listens for 8 seconds. An on-device model detects whether the sound is a cry and checks recording quality; a silent, noisy or short recording is rejected. Context then ranks the reasons: time since the last feed, time awake, the last diaper change, your baby's age, the hour of the day and, once enough examples exist, your “right / wrong” feedback. The result is a ranked list with probabilities; when the possibilities are close the app says “I'm not sure”, and the “What did this result look at?” panel shows which facts it used. Detail: baby cry analyzer.

Responsible use

Agubu is a care tool; it is not a medical device and does not replace a clinician's assessment. The guide articles are general information and have not been reviewed by a clinician; each one lists its sources at the end. If crying lasts more than 3 hours and nothing settles it, if there is fever, vomiting or a rash, if the cry sounds unusually high-pitched or weak, if your baby refuses to feed, or if your instinct says something is wrong, put the app down and see a doctor. In an emergency, call your local emergency number.

Privacy and money

No account, no server, no ads, no analytics. Records stay on your phone; audio is never written to disk. Revenue comes only from the Premium subscription; the free tier is permanent, and basic logging, 3 analyses a day, 8 sounds, the last 7 days of history, safety content and the methodology page are never put behind the paywall. Detail: privacy.

Contact, languages, platforms

Email: ygtsmt2999@gmail.com — every email is read and answered within 7 days at the latest
Languages: English and Turkish (app and website)
Platforms: iPhone (iOS 15.0 or later; on iPad in iPhone compatibility mode) and Android
Available on the App Store and Google Play — download
Support page: support

Frequently asked questions

Is Agubu a company?

No. It is the product of one independent developer in Turkey, Yiğit Samet Ölmez. There is no company, investor or team.

What sets Agubu apart from other baby cry apps?

It does not claim to read the reason from the sound. It detects the cry, ranks the reasons from your baby's day, says when it isn't sure, and publishes every number it measured. And because there is no account and no server, your records never leave your phone.

Were doctors or pediatricians involved?

No. The guide articles draw on the general guidance of bodies such as the AAP, the NHS and the Turkish Ministry of Health and list their sources, but they have not been reviewed by a clinician. We make no claim of such an arrangement.