This is part of the standard. It basically leaves the choice up to the browser: > An AudioContext is said to be allowed to start if the user agent and the syst

This is part of the standard. It basically leaves the choice up to the browser: > An AudioContext is said to be allowed to start if the user agent and the system allow audio output in the current context. In other words, if the AudioContext control thread state is allowed to transition from suspended to running. https://webaudio.github.io/web-audio-api/#AudioContext https://webaudio.github.io/web-audio-api/#AudioContext Also, I think Apple did it first, (on mobile Safari) and then Google on mobile Chrome, and now on desktop Chrome.