Split Any Song Into Stems
Powered by Demucs AI, Gig Friend separates any audio file into vocals, drums, bass, and other instruments — so you can isolate exactly what you need to learn.
Try Stem Splitting Free4-Stem Separation
Vocals, drums, bass, and other — each stem gets its own track with independent volume, mute, and solo controls.
Cloud-Processed
Stems are separated on powerful cloud GPUs using Demucs, so you get studio-quality results regardless of your device.
Inline Playback
Each stem renders its own waveform. Mute the vocals to practice singing, or solo the bass to learn the bassline note-for-note.
Hear every instrument in isolation
Ever struggled to hear the guitar part buried in the mix? Stem splitting lets you strip away everything else and focus on the part you need to learn. Each separated track includes its own visual waveform so you can see dynamics and timing at a glance.
- Mute vocals to create instant karaoke backing tracks
- Solo the drum track to study groove patterns
- Isolate the bass for transcription practice
Typical processing time: 30-90 seconds
Powered by state-of-the-art AI
We use Meta's Demucs model running on cloud GPUs to deliver the highest-quality stem separation available. Upload your track, and within 30-90 seconds you'll have four clean, separated stems ready to play.
Waveform peaks are pre-computed server-side for instant rendering — no waiting for waveforms to load on your device.
How It Works
Upload your song
Drag any audio file (MP3, WAV, FLAC, M4A) into Gig Friend.
Tap "Split Stems"
One click sends your track to our cloud for AI processing.
Stems appear automatically
In under 90 seconds, four separated stems appear with individual waveforms and controls.
Mix and practice
Mute, solo, or adjust volume on any stem. Loop sections to drill difficult passages.
Available on Hobby and Ultimate plans
Hobby includes 10 AI credits/month. Ultimate includes 50 AI credits/month. Each stem split uses 1 credit.
View Pricing