System audio capture lets you record any sound output from your computer. It works with every application — meeting platforms, webinars, training videos, podcasts, or any browser tab. This feature is exclusive to the Mavio desktop app.

​
How it works

The Mavio desktop app installs a lightweight virtual audio driver that mirrors your system’s audio output. When you start a system audio recording, Mavio captures the mixed audio stream before it reaches your speakers or headphones. Your actual audio output is unaffected — you hear everything as normal.
System audio capture records what you hear, not what your microphone picks up. To include your own voice (for example, during a meeting), enable Microphone + System Audio mode.

​
Getting started

1

Install the desktop app

Download Mavio for macOS or Windows and complete the setup process.
2

Grant audio permissions

macOS: Go to System Settings > Privacy & Security > Screen Recording and enable Mavio. A restart of the app may be required.Windows: No additional permissions are needed beyond the initial microphone access prompt.
3

Start a recording

Click the Mavio icon in your menu bar (macOS) or system tray (Windows) and select Record System Audio. Alternatively, use the global hotkey:
  • macOS: Cmd + Shift + R
  • Windows: Ctrl + Shift + R
4

Stop the recording

Click the recording indicator in the menu bar or system tray, or press the hotkey again. Mavio processes the audio and delivers your transcript, summary, and action items within minutes.

​
Recording modes

ModeCapturesUse case
System audio onlyAll audio output from your computerWebinars, training videos, podcasts
Microphone onlyYour microphone inputIn-person dictation, voice memos
System audio + microphoneBoth streams, mixedActive meeting participation where you want your voice included
Switch modes from the recording dropdown in the menu bar or in Settings > Recording > Default capture mode.

​
Audio quality

System audio capture records at the native sample rate of your audio output (typically 44.1 kHz or 48 kHz) in 16-bit PCM. The audio is compressed before upload using Opus codec at 64 kbps, which preserves speech clarity while keeping file sizes small.
For the best transcription accuracy, keep your system volume at a moderate level. Extremely low volume can reduce transcription quality, while clipping from very high volume degrades accuracy.

​
Speaker identification with system audio

Since system audio is a mixed stream (all speakers combined into one channel), Mavio uses AI-based speaker diarization to separate individual voices. This works best when:
  • Speakers take turns (minimal crosstalk)
  • Each speaker has a distinct voice profile
  • Audio quality is clear without heavy compression or background noise
If you have previously identified speakers in past meetings, Mavio will attempt to match voices to known profiles automatically.

​
Platform-specific notes

  • Requires Screen Recording permission (used for the virtual audio driver, not actual screen capture)
  • Works with all audio outputs including built-in speakers, headphones, and external DACs
  • Compatible with macOS 12 Monterey through macOS 15 Sequoia
  • Apple Silicon and Intel Macs are both supported

​
Troubleshooting

Verify that Mavio has Screen Recording permission on macOS (System Settings > Privacy & Security > Screen Recording). On Windows, ensure no other application is exclusively locking the audio device.
Switch to System Audio + Microphone mode in the recording dropdown. System audio alone captures only what plays through your speakers.
Check your system volume — if it is at 100% and the source is loud, audio may clip. Lower the volume to 70-80% and try again. Also ensure no audio enhancement software is modifying the output.