With the widespread use of smartphones that have multiple sensors and sound processing capabilities, there is a great potential for increased audience participation in music performances. This paper proposes a framework for participatory mobile music based on mapping arbitrary accelerometer gestures to sound synthesizers. The authors describe Handwaving, a system based on neural networks for real-time gesture recognition and sonification on mobile browsers. Based on a multiuser dataset, results show that training with data from multiple users improves classification accuracy, supporting the use of the proposed algorithm for user-independent gesture recognition. This illustrates the relevance of user-independent training for multiuser settings, especially in participatory music. The system is implemented using web standards, which makes it simple and quick to deploy software on audience devices in live performance settings.