Health tech company Lumme Health has just released the world’s first smartwatch app to automate calorie tracking — a laborious activity that typically requires looking up the calorie values of individual foods. The kCalculator smartwatch app runs in the background on an Apple Watch and uses hand motions to passively detect eating, and scientifically-proven algorithms to estimate calories.

Similar to other apps that use motion sensor data to detect a particular activity, the kCalculator senses when a user is eating from the motions of the dominant hand and sends corresponding ‘eating detection’ notifications for the user to verify on their phone or Apple Watch. A single ‘Meal Diary’ screen on the phone app displays a running tally of the day’s calories along with the time, duration, pace and calorie estimates of each eating session. Users are free to eat normally using utensils or their hands, and any eating that lasts for at least a few minutes can be detected.

“The #1 question we get is how the app knows what you’re eating,” says Lumme Health Co-Founder and CEO Arun Mathur, who’s no stranger to the challenges of educating consumers on innovative technologies. Twenty years ago, he stood in front of his red DVD rental machines in New York City explaining the benefits of automating movie rentals. The business soon caught on and was acquired by Redbox.

“The question itself speaks to the painstaking nature of traditional calorie counting and also why it’s so highly prone to error,” says Mathur.  “What you eat will always matter to good health, but the math on calories depends a lot on portion sizes – simply how much you eat. Portion size tends to get vastly underestimated, but it’s something we measure pretty well in the form of a bite. By solving for volume first, adjusting for what you eat is surprisingly simple and effective.”

The kCalculator’s bite-based algorithms use a personalized caloric proxy for each bite based on an individual’s profile details (gender, age, height, weight) and allow for calorie adjustments using a ‘Calorie Type’ slider (low/mix/high) that captures the vast majority of variability in what a user eats.

The real breakthrough comes from the company’s patented technology, which uses machine learning to sift through tens of thousands of daily hand movements to find the 100+ bites of food a person might actually take. It’s also required to learn the unique eating gestures of the individual, making the app more personalized with daily use and more intelligent in its detection accuracy and resulting calorie estimates.

Lumme’s Co-Founder and Chief Behavioral Scientist, J. Graham Thomas, PhD, who brings over 15 years of hands-on clinical research experience in psychiatry and obesity medicine says, “Journaling everything we eat to achieve a calorie goal can be a tedious chore that too often uses up energy that could be directed towards mastering the actual habits that matter to long-term success. For the first time, Lumme’s technology removes a critical area of friction in weight management, which paves the way for more effective goal setting, progress tracking, and problem-solving. Calorie tracking scratches the surface.”

The kCalculator is a slice of the company’s forthcoming behavioral app for weight loss, which provides actionable insights, daily predictions and real-time interventions to help users achieve their goals and avoid cycles of problematic behavior.

The kCalculator A.I. Calorie Counter for Apple Watch (Series 4/5/6/7/SE) is available as a FREE download on the App Store for a limited time. For more information, visit lummehealth.com.