SystemAudioKitAPI reference
Every public type in SystemAudioKit.
AudioRecorder
public final class AudioRecorder: records the microphone and system audio together.
| Member | Description |
|---|---|
init() | A new recorder. |
start(_ configuration:, in directory:) async throws | Starts recording into directory, writing microphone.wav and/or system.wav. |
stop() async -> Recording? | Stops and returns the finished tracks, padded to the same length. nil when nothing was recording. |
isRecording: Bool | Whether a recording runs. |
elapsed: TimeInterval | Seconds since start. |
onLevels: ((Float, Float) -> Void)? | Live microphone and system levels, 0 to 1, called on audio threads. |
systemAudioPeak() -> Float? | Loudest system level since the last call, then resets. nil without a system track. |
AudioRecorder.Configuration
Prop
Type
Recording
A finished recording. Sendable, Equatable.
Prop
Type
mix(to url: URL) throws: mixes all tracks into one.wavor.m4a. See Mixing.
SystemAudioSource
.everything: everything the Mac plays, except this app..apps([String]): only these bundle IDs, including their helper processes..everythingExcept([String]): everything except these apps and this app.
SystemAudioBackend
.processTap: Core Audio process tap. Default. Supports app selection..screenCaptureKit: display audio through ScreenCaptureKit. Needs Screen Recording and always records everything.
TrackFormat
sampleRate: Double (default 16 000) and channels: AVAudioChannelCount (default 1). TrackFormat.speech is the default. Files are written as 16-bit PCM WAV.
AudioPermissions
| Member | Description |
|---|---|
microphone, systemAudio, screenRecording | Status: .authorized, .denied, .notDetermined or .unknown. |
requestMicrophone() async -> Bool | Shows the microphone prompt. |
requestSystemAudio() async -> Bool | Shows the system audio prompt. Uses private TCC calls. |
settingsURL(for: Pane) -> URL | Opens .microphone, .systemAudio or .screenRecording in System Settings. |
See Permissions for the caveats of the system audio check.
MeetingDetector
| Member | Description |
|---|---|
init(watchedApps:startDelay:endDelay:) | Defaults: knownApps, 4 s, 8 s. |
static knownApps: [String: String] | Bundle ID to name for common call apps and browsers. |
onChange: ((started, ended) -> Void)? | Called on the main queue. |
active: [DetectedMeeting] | Calls in progress. |
start(interval:), stop() | Polling on the main run loop. Default interval 2 s. |
currentMeetings() -> [DetectedMeeting] | Apps using the microphone right now, without delays. |
DetectedMeeting: bundleID, appName, processIDs.
AudioDevices
| Member | Description |
|---|---|
inputs(), outputs() -> [AudioDevice] | Physical devices. Virtual and aggregate devices are left out. |
input(uid:) -> AudioDevice? | The connected input with this UID. |
defaultInputID, defaultOutputID | Core Audio IDs of the defaults. |
AudioDevice: id, uid (stable, store this one), name, isDefault, transport (.builtIn, .usb, .bluetooth, .virtual, .aggregate, .other).
AudioProcesses
| Member | Description |
|---|---|
all() -> [AudioProcess] | Every process Core Audio knows. |
playing() -> [AudioProcess] | Processes producing output now. |
objectID(for pid:) -> AudioObjectID? | The Core Audio object of a process. |
objectIDs(forApp bundleID:) -> [AudioObjectID] | An app's processes, including helpers. |
AudioProcess: id, pid, bundleID, appBundleID (helpers followed to their app), name, isRunningInput, isRunningOutput.
AudioMixer
mix(_ inputs: [URL], to output: URL) throws: sums tracks with 1/√n gain and soft clipping.AudioMixer.Error:.noTracks,.unsupportedExtension(String).
Lower level
AudioRecorder is built from parts you can use directly:
| Type | Description |
|---|---|
AlignedTrackWriter | Writes buffers to a WAV file on a host-time timeline, filling gaps with silence. init(url:format:startHostTime:), append(_:hostTime:), finish(at:), duration, takePeak(), onLevel. |
ProcessTapCapture | A process tap feeding a writer. start(source:writer:), stop(), lastBuffer. |
MicrophoneCapture | A microphone feeding a writer. start(deviceUID:writer:), stop(). |
ScreenCaptureAudioCapture | ScreenCaptureKit audio feeding a writer. start(writer:excludingOwnAudio:), stop(). |
HostClock | now (mach_absolute_time) and seconds(from:to:). |
AudioLevel | normalized(_:): RMS of a buffer mapped from −50 to 0 dBFS onto 0 to 1. |
Give every writer of one recording the same startHostTime, so their sample 0 is the same moment.
Errors
SystemAudioError:
.microphoneUnavailable: no input device..microphonePermissionDenied,.systemAudioPermissionDenied..noMatchingProcesses([String]): none of the apps in.appshas an audio process..noDisplay: ScreenCaptureKit found no display..tapFormatUnavailable: the tap's audio format could not be read..alreadyRecording.
CoreAudioError carries the OSStatus and the operation that failed.