Can QuickTime record system audio? No — here’s why, and what to use

No. QuickTime Player’s audio recorder offers exactly one input control, and it is a microphone list. There is no system-audio setting buried in a menu, no defaults-write incantation, and no version of QuickTime that has ever done this.

That is the whole answer. The rest of this page is the proof, and then what to do instead.

Apple’s own wording, in three places

The fastest way to stop looking is to read Apple’s documentation for each surface people try. Here is what each one lists as its audio input.

QuickTime Player, New Audio Recording. Two controls, quoted verbatim:

Microphone: Choose a microphone (if you have more than one available).

The only other control is Quality:

Choose the recording quality. Maximum-quality recordings produce uncompressed files, which can use large amounts of storage space.

Apple goes into enough detail to tell you that Maximum means “44100 Hz, uncompressed Linear PCM 24-bit AIFF-C audio”. A page that specific about bit depth is not omitting a system-audio option by accident.

The screenshot toolbar, ⌘⇧5. Same story:

Below Microphone, select the microphone you want to use so that a checkmark appears next to it. To turn off the microphone, select None.

Microphone, or nothing. This is why so many people end up with a perfect video of a call and silence where the other person was.

Voice Memos.

To create a recording, use the built-in microphone, a supported headset, or an external mic.

Three input paths, all of them microphones.

macOS can do it. Apple just doesn’t offer it to you

Here’s the part that makes the limitation feel arbitrary until you see the reason for it. The capability exists in macOS, and Apple documents it — as a privacy control for other people's apps. The panel is called Screen & System Audio Recording, and Apple’s description of it is:

You can allow apps to record both your screen and audio, or just your audio.

“Just your audio” is the sentence that matters. macOS has a real, permissioned, user-revocable channel for capturing the sound your Mac is playing. Apple built the gate, shipped the switch, and then declined to put a system-audio checkbox in QuickTime.

The logic holds up, even if it’s annoying. Your microphone records you and your room. Your Mac’s output can contain a video call, a voice message, someone’s phone call, a film you paid to watch — audio belonging to people who never agreed to be recorded. Apple treats that as a separate class of access, gated behind a prompt and a list you can audit. What it doesn’t do is put that power in a first-party app where it would be one click away by default.

Route 1 — give QuickTime a fake microphone

Since macOS hands every audio input device to every app, you can manufacture an input that is really your output. That’s what a virtual audio driver does.

BlackHole is the usual choice. It’s free with a suggested $10 donation, and its own page states the compatibility plainly: “Compatible with macOS Mavericks (10.10) and newer” and “Built for Intel and Apple Silicon”, in 2-, 16- and 64-channel builds. Loopback by Rogue Amoeba is the paid alternative with a graphical patchbay instead of Audio MIDI Setup.

The shape of the setup: install the driver, build a Multi-Output Device in Audio MIDI Setup containing both your real speakers and the virtual device, switch your Mac’s output to it, then pick the virtual device as QuickTime’s input. Skip the Multi-Output step and you record fine but hear nothing.

What it costs you, honestly:

• A Multi-Output Device ignores the volume keys. You lose keyboard volume control while it’s selected.
• Getting your own voice in as well means stacking an Aggregate Device on top, and everything lands mixed into one track.
• An aggregate device does not mix its channels for you — it hands the app N separate channels, which QuickTime won’t split. This is a common route to “only the left side has sound”.
• Every recording begins with remembering to switch and ends with remembering to switch back.

For a one-off capture — pulling audio out of a video, keeping a stream — this is completely fine, and free. As a daily ritual before every meeting it is the actual problem.

Route 2 — an app that holds the permission

The alternative is to skip the plumbing. An app can request the system-audio permission and read the output stream where it already is. Your sound settings never change, your headphones keep working, and there is nothing to undo afterwards.

Narada works this way, and two things follow from it.

Your microphone and the call stay in separate files. They’re captured as two independent streams, then transcribed separately. Your own lines are attributed to you because they came out of your own microphone. The other side gets split into distinct voices, labelled Speaker 1, Speaker 2 and so on — the app has no way to learn anyone’s name and doesn’t invent one. Those labels come from the call’s audio, so a mic-only recording of a room has nothing to separate and won’t get them.

It never asks for Screen Recording. System audio arrives through a Core Audio process tap rather than a screen-capture API, so the permissions are Microphone and System Audio Recording. Your screen is not involved.

Transcription runs on the Mac with Parakeet TDT v3 through Core ML, across 25 European languages. Accuracy is not uniform and it would be dishonest to publish one number: on the model vendor’s own FLEURS benchmark, word error rate runs from about 4% for Italian to about 37% for Greek. Narada needs macOS 15.6 or later. Recording, transcription and the plain-file archive are in the free tier.

One more thing, since this page is about a tool that quietly gives you less audio than you expected: macOS audio can fail mid-recording. A device can stall, the HAL can drop cycles, a microphone can go dead without notifying anyone. Narada compares each finished recording against what was submitted to the file and marks it Incomplete, with a reason, when the numbers disagree. It cannot get the missing audio back. Nothing can. What it can do is stop a half-empty file from being presented to you as a complete meeting.

Which route for which job

One thing no software settles for you

If what you’re recording has another person in it, tell them. Recording rules differ by country and by state, and some places require every participant’s consent rather than just yours. No app can grant you permission, and no app notifies the room on your behalf — Narada keeps a reminder on screen while it records, and that is a reminder, not consent. Saying it out loud at the start takes four seconds.

Sources

• Apple Support, Record audio in QuickTime Player on Mac — the Microphone and Quality controls, and the AIFF-C detail. https://support.apple.com/guide/quicktime-player/record-audio-qtpf25d6f827/mac
• Apple Support, Take a screenshot or screen recording on Mac — “Below Microphone, select the microphone you want to use”. https://support.apple.com/guide/mac-help/take-a-screenshot-or-screen-recording-mh26782/mac
• Apple Support, Create a recording in Voice Memos on Mac — “the built-in microphone, a supported headset, or an external mic”. https://support.apple.com/guide/voice-memos/create-a-recording-vmaa4b813415/mac
• Apple Support, Control access to screen and system audio recording on Mac — the panel name and “both your screen and audio, or just your audio”. https://support.apple.com/guide/mac-help/control-access-screen-system-audio-recording-mchld6aa7d23/mac
• Existential Audio, BlackHole — free with suggested donation, “Compatible with macOS Mavericks (10.10) and newer”, “Built for Intel and Apple Silicon”. https://existential.audio/blackhole/

All five checked on 4 August 2026.