Why macOS still doesn't ship a per-app volume mixer in 2026

Microsoft shipped a per-app volume mixer in Windows Vista in 2006. PulseAudio added per-app volume to Linux around the same time. Twenty years later, macOS still ships with one slider for the whole system.

Why? A short, uncharitable history.

Phase 1: Core Audio's clean API (2001–2010)

OS X's audio stack, Core Audio, was a marvel when it shipped — clean, low-latency, designed for pro audio. But it was built around the device as the unit of audio, not the application. Apps requested an output device and pushed samples to it. There was no system-level concept of "this app is producing audio" that the OS could attenuate independently.

If you wanted per-app volume, you needed something sitting between the app and the device. That something had to ship as a kernel extension or a virtual driver. Third parties (Rogue Amoeba, others) built exactly that. Apple itself never did.

Phase 2: "Mac just works" (2010–2019)

Through the 2010s, Apple's strategy was clearer: the Mac was a media-consumption device for most users, and "the audio just works" was the user-facing promise. A per-app mixer is a power-user feature. Adding it to System Preferences would expose a confusing surface to people who didn't need it.

This is the same logic that kept the right-click menu out of macOS for years. Apple's bet was that exposing too many controls would make the experience worse for the median user. So third parties got to fill the gap, and Apple stayed out of it.

Phase 3: the kernel extension lockdown (2019–2023)

Then macOS Catalina happened. Apple deprecated kernel extensions and pushed everyone toward DriverKit and signed system extensions. The audio tools that had quietly powered the per-app mixer category for a decade had to be rebuilt. Many didn't survive — BackgroundMusic, for example, lost active maintenance for a stretch.

For a few years, the answer to "can I get a per-app mixer on macOS" became "yes, but the install process is going to scare you and it might not survive the next OS update."

Phase 4: Sonoma's quiet shift (2023–now)

macOS Sonoma added per-process audio APIs that finally let third parties read and influence individual apps' audio streams without virtual drivers or kernel extensions. It wasn't a marquee feature — Apple buried it in a Core Audio release note. But it's the API mactooloud is built on, and it's why a new generation of audio utilities can finally exist without an install ritual.

Will Apple ever ship its own?

Probably not in a year that ends in 6. Apple doesn't ship features just because Windows has them; they ship them when they have a coherent design opinion. Per-app volume isn't a feature Apple has ever shown design interest in, and the workarounds (System Preferences → Sound, Audio MIDI Setup, third parties) work well enough that the pressure isn't there.

So we shipped mactooloud. If Apple ever changes their mind, we'll happily get out of the way.