Like layers of Tiramisu, we use AI to separate your music into vocals, drums, bass, and instruments
100% Free • Open Source • Native macOS App
Cocoa Dust
Ladyfinger Grid
Espresso Pour
Mascarpone Cream
Four perfectly separated layers, like a tiramisu
Uses state-of-the-art Demucs AI models from Facebook Research to separate stems with incredible accuracy
Leverages Metal Performance Shaders for blazing-fast processing on Apple Silicon and Intel Macs
Choose between Fast Mode (2-4 min) for quick results or Premium Mode (6-10 min) for maximum quality
Native macOS design with a clean, modern interface that feels right at home on your Mac
Supports MP3, WAV, FLAC, M4A, AAC, and OGG input files. Outputs high-quality WAV stems
Completely free to use with no restrictions. Open source under MIT license on GitHub
Select any audio file from your library - songs, podcasts, or any audio recording
Pick Fast Mode for quick separation or Premium Mode for maximum quality
Our AI analyzes your music and intelligently separates it into distinct layers
Receive vocals, drums, bass, and instruments as separate high-quality audio files
Free for macOS • Version 1.0
macOS 10.15 (Catalina) or later
~613 MB (includes AI models)
If you see a security warning when opening Tiramisu, go to System Preferences → Security & Privacy and click "Open Anyway". This is normal for apps distributed outside the Mac App Store.
Create mashups, remixes, and acapellas by extracting vocals and instrumentals
Practice along with isolated instrument tracks or create backing tracks
Extract vocals for covers or remove vocals to create karaoke versions
Analyze arrangements by listening to individual stems and learning from the pros
Tiramisu is completely free and open source. If you find it useful, star the project on GitHub and share it with others!
Tiramisu is built by the community, for the community. Share it with friends who love music!