SystemAudioDump

Recall.ai - API for desktop recording

If you’re looking for a hosted desktop recording API, consider checking out Recall.ai, an API that records Zoom, Google Meet, Microsoft Teams, in-person meetings, and more.

This project is sponsored by Recall.ai.

---

[!NOTE]
This tool captures SYSTEM AUDIO only (apps, browser, media players, etc.) and does NOT record from external sources like microphones, external audio interfaces, or USB devices

A macOS command-line tool that captures system audio and outputs it as raw PCM data to stdout.

Features

• Captures all system audio (excluding the current process)
• Converts audio to 24kHz, 16-bit PCM format
• Real-time streaming to stdout
• Automatic permission handling

Requirements

• macOS 13.0+ (for ScreenCaptureKit support)
• Xcode with Swift 5.7+
• Screen Recording permissions

Building

swift build -c release

Usage

Run the executable:

./.build/release/SystemAudioDump

The tool will:

1. Check for screen recording permissions (required for system audio capture)
2. Prompt you to grant permissions if needed
3. Start capturing system audio and output raw PCM data to stdout

Redirecting Audio Output

Pipe the output to a file or another program:

# Save to raw PCM file
./.build/release/SystemAudioDump > audio.pcm

# Play through ffplay
./.build/release/SystemAudioDump | ffplay -f s16le -ar 24000 -ac 2 -

Permissions

When first run, macOS will prompt for Screen Recording permission. This is required because system audio capture uses the same privacy framework as screen recording.

Go to: System Preferences > Security & Privacy > Privacy > Screen Recording and enable access for your terminal or the application.

Output Format

• Sample Rate: 24kHz
• Bit Depth: 16-bit signed integers
• Channels: Stereo (2 channels)
• Format: Interleaved PCM, little-endian

Stopping

Press Ctrl+C to stop the capture and exit gracefully.