Swigable set out to build a premium smart water bottle that helps people drink enough: track what you actually drink, learn your pattern, nudge you at the right moment. The hard part is the first step. A bottle knows how it is tilted, and that is all it knows. Tilting to drink looks a lot like being knocked over, swung in a bag or laid down on its side, and a hydration tracker that counts those is a toy. Swigable wanted the real thing, and came to us to build it.
Machine learning that tells a real sip from a bottle rolling around in a bag, built on academic data, inside a premium smart bottle launching soon.
01
The start
Counting sips is harder than it sounds

02
The journey
The algorithm, the app and the hardware
The measurement runs on machine-learning models built from academic data on how people drink, so the bottle judges the volume of a real sip from tilt alone, with a high degree of accuracy, and rejects everything that is not one. Around that we designed the product: a mobile app [iOS and Android] that tracks in real time, adapts to the user’s pattern and nudges when a drink is due, with Swiggy, the in-app hydration coach; and the hardware itself, a waterproof magnetic health device that snaps onto a stainless steel bottle, which we designed and specified for the manufacturers.


03
The result
Launching soon
Swigable launches soon.
- Real-time sip tracking from bottle tilt alone
- Machine learning built on academic data on drinking behaviour
- Rejects false readings from movement, knocks and falls
- An app that learns your pattern and nudges at the right moment
- Hardware designed and specified by us for manufacture
- Launching soon


