Why macOS has no volume mixer.

Windows has had a per-app volume mixer built into the OS since Vista, in 2007. macOS, nearly two decades later, still gives you one system-wide volume key. Turn it down, and you turn down Spotify, Slack's notification chime, and your Zoom call all at once. This isn't an oversight Apple is quietly working on — it's a consequence of how Core Audio was designed, and it only recently became possible for a third-party app to fix.

Core Audio doesn't route audio through apps you can intercept

On Windows, the audio session API was built with per-application mixing as a first-class concept from the start. On macOS, Core Audio mixes every app's output into a shared path down to the hardware, and for most of macOS's history there was no public, sanctioned way for a third-party app to sit between an individual process and that path — only the system itself could see and adjust it.

The workaround the community landed on was a virtual audio driver: an app installs a fake audio device, routes your Mac's system output into it, and then remixes what comes out the other side. Soundflower (2004) pioneered this; Rogue Amoeba's Loopback and older versions of SoundSource, along with BackgroundMusic, followed the same pattern for years. It worked, but it meant a kernel- or system-level driver install, and on Apple Silicon Macs, approving that driver typically meant a restart into Recovery Mode to lower a security setting — a real barrier for anyone who wasn't already comfortable in System Settings.

What changed: Core Audio process taps

Apple shipped a new, public Core Audio API — process taps — starting in macOS 14.2 (Sonoma, released December 2023), with the tooling third-party developers actually rely on maturing by macOS 14.4 (March 2024). A process tap lets a signed app capture the audio a specific process is producing, without a virtual driver and without installing anything at the system level. Attach it to an aggregate device, and you can adjust the gain, run it through a filter, or redirect it entirely — all in user space, with macOS gating access behind a one-time System Audio Recording permission prompt rather than a kernel extension approval flow.

This is the API MixPill is built on, and it's the same API behind most of the current generation of per-app mixers. It's still fairly new and Apple's own documentation on it is sparse — which is part of why implementations vary in polish.

The options, as of 2026

Fairly, and without pretending MixPill is the only one:

• SoundSource (Rogue Amoeba) — the long-standing, most established option in this category, now on version 6 and requiring macOS 15+. A one-time $49 purchase, per-app volume and routing, and a deep feature set built over many years.
• SoundDial (€14.99 one-time) and Submix (one-time purchases starting around $9.99) — newer, lower-priced mixers covering the same driver-free territory.
• MixPill — free for unlimited per-app volume, mute and metering; Pro adds EQ, output routing and Auto-Ducking. See the pricing for specifics.

Where MixPill diverges from the rest is what's free: per-app volume, mute and live metering cost nothing, for unlimited apps, forever — no trial, no nag. Pro adds per-app EQ and output routing on top, plus, unlike anything else in this list, a DAW that's never tapped and background audio that ducks itself the moment a call starts.

The honest caveat

Because process taps are relatively new and thinly documented, every app in this category — MixPill included — is still building on ground Apple hasn't fully paved. Expect rough edges to surface with new macOS releases faster than you would with older, more settled APIs.