Stem Studio

What Stem Studio does

Drop any audio file and let the on-device AI model break it down into its individual musical components — all processed privately on your Mac.

🎙️

6-Stem Separation

Powered by HTDemucs v4, Stem Studio isolates Vocals, Drums, Bass, Guitar, Piano and an Other track with professional-grade accuracy.

🖥️

100% On-Device

All AI inference runs locally using Apple's Neural Engine. Your audio never leaves your Mac — no internet connection required after installation.

⚡

Apple Silicon Optimised

Built with Core ML and Metal acceleration. A 3-minute song is separated in roughly 60–90 seconds on an M2 chip.

📁

Project Management

Organise your work in named projects. Each project stores the original file, all separated stems and export history so you can pick up where you left off.

🎛️

Individual Playback

Preview every stem in isolation directly inside the app. Solo the guitar, mute the drums, or loop any section to check the separation quality before exporting.

💾

WAV Export

Export each stem as a high-quality WAV file to any folder on your disk. Full-resolution audio — ready for your DAW of choice.

How it works

Three steps from song to stems.

Drop your file

Drag an MP3, WAV, AIFF or AAC file onto Stem Studio or pick one from the menu. The file is loaded instantly.

AI processes it

The HTDemucs model runs on your Neural Engine. A progress indicator shows you which stems are being extracted in real time.

Export & enjoy

Listen to each stem in the built-in player, then export them all as individual WAV files in one click.

System Requirements

🍎

macOS 14 Sonoma or later

Stem Studio is a native Mac app built with SwiftUI. macOS 14 Sonoma is the minimum supported version.

💻

Apple Silicon recommended

M1, M2 or later chips deliver the fastest separation times via the Neural Engine. Intel Macs are also supported but will be slower.

🔒

No internet required

Once installed, Stem Studio works entirely offline. Your audio files stay on your device at all times.