You hit record on QuickTime, walk through the demo, and when you play the file back there’s no sound from the app you were showing. Just your own voice from the microphone, or dead silence if you didn’t have one plugged in. This happens to nearly everyone who tries to record a tutorial, a bug report, or a song playing back on a Mac for the first time, and it isn’t a bug in QuickTime. It’s doing exactly what it was built to do.
Why QuickTime never hears the app
QuickTime’s screen recording, and the Cmd+Shift+5 toolbar underneath it, is built to open the microphone as an audio input. That’s the only audio source either one knows how to reach. Neither has a way to read the mix that macOS sends to your speakers, because an app listening in on another app’s output isn’t something Core Audio allows by default. If it were, any app could quietly record whatever you’re hearing, including a call.
Windows sound cards have often shipped with a hardware loopback input, usually called Stereo Mix, that routes the output signal back in as a recordable source. macOS doesn’t expose an equivalent out of the box. To capture the sound that actually reaches your speakers, something has to sit between the apps and the output device and duplicate that signal into a form a recorder can open. Apple doesn’t ship that piece. It has to come from a driver.
What actually works
The oldest trick is a physical loopback: a cable from the headphone output back into the microphone input. It works, technically, but the signal goes through an extra digital-to-analog and analog-to-digital pass, picks up whatever noise is in the room, and adds latency that drifts out of sync with the video over a long recording. Worth knowing it exists, not worth doing if you have another option.
BlackHole is free, open source, and genuinely solves the problem. It installs a virtual audio device; you build a Multi-Output Device in Audio MIDI Setup that sends sound to both BlackHole and your real speakers, then pick BlackHole as the recording source in QuickTime. The trade-off is setup: getting mic and system audio onto separate, mixable tracks means building an aggregate device too, and there’s no per-app control, it’s the whole system mix or nothing.
OBS Studio is also free, and on current macOS its audio output capture source uses Apple’s own ScreenCaptureKit to grab system audio without installing a separate driver at all. For anyone comfortable with a scene-based recorder, this is probably the strongest free option available right now, and it’s worth trying before reaching for anything paid.
Paid dedicated recorders like ScreenFlow, Camtasia, Loom or CleanShot X handle the routing for you and add editing on top, at the cost of a subscription or a one-time license. If you only need this occasionally, the free route is worth trying first; if screen recording is part of your regular workflow, paying for something with an editor built in usually saves more time than it costs.
Recording system audio with ZonaReq
ZonaReq’s HAL driver already sits in the position this problem needs filled: it captures the full system mix for the equalizer, so the same tap point can feed a recording. Screen capture uses ScreenCaptureKit for video, and pulls system audio and microphone from the driver as separate tracks, so there’s no Multi-Output Device or aggregate device to build by hand.
- Install ZonaReq and approve the driver install. It needs an administrator password and restarts the system audio service once, which briefly mutes everything and closes any open call.
- Open the recording module and pick the screen or window to capture.
- Turn system audio and microphone on independently, so either can be muted later without re-recording.
- Record, then trim, split or join clips in the built-in editor before exporting.
A limit worth knowing before you start
The free tier caps recordings at three minutes and adds a watermark to the exported video; Pro removes both. And installing any audio driver, ZonaReq or BlackHole, briefly interrupts system audio for everyone using the Mac at that moment, so it’s not something to do seconds before a live call. If all you need is your own voice narrating a screen, the built-in Cmd+Shift+5 recorder already does that job well, and adding a driver would be solving a problem you don’t have.