What is system audio capture?
System audio capture records the audio coming out of your speakers during a meeting: every other voice on the call. Wispr Notetaker pairs that with your microphone, so your words and everyone else's both land in the transcript.
What system audio capture means in a meeting tool
Every voice on a video call arrives at your machine as playback. Your speakers or headphones turn it into sound; the operating system knows it as an audio stream. System audio capture taps that stream directly.
Your microphone is the other half. It carries your voice, plus the room you're sitting in. Put both together and a recording app has the whole conversation: you from the mic, everyone else from system audio.
This is why the term shows up in permission screens rather than marketing pages. Recording what a computer plays is a privileged action on macOS, so the operating system gates it behind an explicit grant. Wispr Notetaker asks for it during setup as "Capture what others say," alongside the microphone permission labeled "Capture what you say." Without system audio, a recording still happens, but it is built from mic audio alone.
System audio capture vs mic-only vs a meeting bot
There are three ways a tool gets a meeting into a transcript: system audio plus microphone, microphone only, or a bot that joins the call. They differ in what the tool hears and in what the room sees.
Wispr Notetaker takes the first row. Nothing joins the call, so the meeting looks normal and recording happens on your device, but you should still tell people you're recording.
Why capture quality decides what your notes are worth
Almost nobody reads the transcript end to end. People read the follow-up email, or go looking for one answer a month later. All of it is assembled from the words in the transcript and the names attached to them.
When your mic picks up remote participants through your speakers, one person can occasionally split into two anonymous speakers. When Wispr Notetaker captures remote participants directly from system audio instead, each voice arrives as its own stream. Named speaker attribution is built on that. It knows who is in the meeting from the calendar invite, knows your names and terms from your personal dictionary, and works out the rest from the conversation.
Get a name wrong and the follow-up goes to the wrong person. Get a term wrong and the customer quote is unusable.
One meeting, end to end with Wispr Notetaker
Say you have a Google Meet with three people, everyone in their own window. System audio carries the other two voices into the recording; your mic carries yours. During the call the live transcript shows you and them. Afterward the refined transcript names each speaker. If it cannot place someone, you name them once and the name applies across the whole transcript.
Because it records on your device rather than joining through a platform's bot, it works wherever the call happens: Google Meet, Zoom, Microsoft Teams, Webex, Slack huddles, and most others. In-person meetings work too, using one shared mic. People show up as Speaker 1 and Speaker 2 until the refined transcript sorts them out.
Named speakers work in fewer places. On Mac, that means Google Meet in Chrome or Edge, Zoom, and Slack huddles in the Slack desktop app. Each one needs Accessibility permission turned on. Wispr Notetaker supports English today, with more languages coming soon.
Frequently asked questions
Does system audio capture record the other person's microphone?
Not their microphone directly. It records what your computer plays, which is the audio the meeting app sends to your speakers. That includes every remote participant, and it also includes anything else playing at the time, such as a video someone shares during the call.
Is system audio capture the same as screen recording permission?
On macOS they live in the same place. The setting sits under Privacy and Security, in the Screen and System Audio Recording pane, so granting audio looks like granting screen access. Wispr Notetaker uses the audio side of it to hear other participants, and asks for it as "Capture what others say" during setup. It opens the System Audio Recording pane so you can toggle Wispr Notetaker on there.
What happens if the system audio permission drops mid-meeting?
Recording keeps going on your microphone alone, and Stop still works. Remote audio from that stretch isn't captured retroactively, though your own mic audio is always preserved. Re-enable the permission, restart the app, and start a new recording. It's worth checking that part of the transcript for gaps.
Can I use system audio capture on Windows?
Not yet. The permission described here is a macOS setting, so on Windows there is no equivalent grant to turn on today. Windows support is coming; until it lands, the capture behavior on this page applies to Mac only.
Does Wispr Notetaker keep recordings of my meetings?
Your meetings are yours. Meeting audio is encrypted and kept only temporarily to create your transcript, let you resume a meeting, verify quality, and troubleshoot. After that limited period, it's automatically removed. Wispr Notetaker doesn't create voiceprints or biometric profiles of you or anyone on your call.
