malahov.io

Record system audio: the whole computer, not one tab

How to capture the audio of any application — desktop Zoom and Teams, a media player, a second browser — and what to switch on in Chrome and macOS first.

Requires version 1.10.0 or newer · Updated 2026-09-10

By default Call Copilot listens to the audio of the tab you are on. Since version 1.10.0 it has a second source — system audio: the extension records everything your computer plays, including applications outside the browser. Desktop Zoom and Teams, a call in a messenger, a local video file, a second browser — all of it reaches the transcript exactly the way tab audio does.

Your microphone is captured alongside it and configured separately; it does not conflict with system audio.

What you need

  • Chrome 141 or newer (or another Chromium browser — Edge, Brave).
  • macOS 14.2 or newer, Windows, or ChromeOS. On Linux Chrome does not hand over system audio, so only tab recording is available there.
  • Call Copilot 1.10.0 or newer. You can check the version in chrome://extensions; updates arrive from the Chrome Web Store automatically.

Step 1. Open "Audio sources"

In the side panel (or in the extension popup) there is a pill under the record button showing the current source. Click it to open the Audio sources panel.

The top section, What to record, is what decides where the audio comes from.

Step 2. Pick "System audio"

Once selected, the pill under the record button reads System audio — that is your confirmation that the next recording will take the audio of the whole system.

The source cannot be switched while a recording runs. Choose it before you press "Start Recording".

Step 3. Start recording and grant access

When you press record, Chrome shows its own picker: a screen, a window, or a tab. Choose what you need and — this is the part people miss — turn on the "Share system audio" toggle at the bottom of the dialog. Without it Chrome hands over a picture with no sound, and the extension stops the recording with exactly that message.

The picture itself is never stored: the video track is stopped immediately and only the audio goes into the transcript.

The picker has a one-minute window. If no screen is chosen by then, the recording does not start — just press "Start Recording" again.

Step 4. macOS only: screen-recording permission

The first time round, macOS asks for permission for Chrome itself. If recording does not start and you see a message about the system denying access, open System Settings → Privacy & Security → Screen & System Audio Recording and enable Chrome in the list. Chrome has to be restarted afterwards.

How to stop

Stop the recording with the button in the extension, the same way you stop a tab recording. The transcript is saved, and summaries and questions about the conversation are available right away.

Chrome's own "Stop sharing" button (in the "You are sharing your screen" bar) ends the recording completely rather than pausing it: a browser cannot resume a screen share that was stopped. The extension closes the session cleanly at that moment and keeps everything recorded so far.

What changes in system-audio mode

  • The recording does not belong to a particular tab, so tab-bound controls — muting the tab and moving the recording to another tab — are hidden.
  • The recording survives closing the side panel and switching tabs: it runs in the extension's background document.
  • Billing is unchanged: one minute of transcription is one credit, whatever the source.

If something does not work

What you seeWhat it means
"picker cancelled"The screen picker was closed without a choice. Press record again.
A message about permission in System SettingsmacOS is not letting Chrome record the screen — grant access (step 4) and restart the browser.
"no audio in the shared source"A screen was chosen, but the "Share system audio" toggle stayed off.
"isn't available in this browser"The browser or the system does not support system-audio capture — check the versions under "What you need".
Silence in the transcriptCheck that the sound really goes to the output device being captured, and that the volume is not at zero.