Timely determination of the time of the onset of a heart attack can save a person’s life by ensuring timely call of medical workers. Researchers from the United States have created a system that provides early prediction of an impending heart attack by altered human breathing. The most characteristic sign of an approaching attack may be the appearance of agonal breathing. It is the change in the frequency and depth of breathing that analyzes the algorithm proposed by experts from the University of Washington, and on the basis of preliminary training, predicts an increased risk of a heart attack and signals the need to take urgent measures. In the future, the algorithm can be integrated with voice assistants and smartphones, which will provide the ability to quickly respond to problems and call qualified specialists in case of heart problems. Diagnosis using the algorithm takes only 10-20 seconds, and the accuracy of recognition of agonal breathing is 97%.