Introduction

SystemAudioKit is a Swift package for call and meeting recorders on macOS. It records the microphone and the system audio (what the Mac plays) as separate WAV files on one timeline, so sample N of each file is the same moment.

import SystemAudioKit

let recorder = AudioRecorder()
try await recorder.start(.init(), in: folder)   // microphone + everything the Mac plays
// …
let recording = await recorder.stop()           // microphone.wav + system.wav

Why separate tracks

A single mixed file loses the one thing a recorder knows for sure: which voice came from the microphone. With two aligned tracks, a transcriber can label everything on the microphone as you and only has to tell the other participants apart. ScribeKit does exactly that.

What is inside

AudioRecorder: starts and stops the microphone and system audio together.
Core Audio process taps (macOS 14.2+) for system audio: everything the Mac plays, only chosen apps, or everything except some. No virtual audio driver, and the lighter "System Audio Recording Only" permission instead of Screen Recording.
ScreenCaptureKit fallback that writes the same aligned output.
AlignedTrackWriter: converts audio to one format and fills gaps with silence, so tracks stay in sync even when an app goes quiet.
MeetingDetector: notices when Zoom, Teams, Meet in a browser and other call apps start and stop using the microphone.
AudioMixer: mixes the tracks into one M4A or WAV.
Permission checks, device and process listing, and a sysaudio command for testing.