Skip to main content
Gridshift integrates third-party AI models for source separation. The specific algorithms may change over time as better models become available.

Models

ModeModelOutput Stems
Song StemsHTDemucsVocals, Drums, Bass, Other
Drum ElementsDrumSepKick, Snare, Cymbals, Toms
Models are downloaded on first use. All processing runs locally — no audio is uploaded.

Usage

1

Select an audio clip

2

Open the Command Palette and search

Open the Command Palette (⌘K) and search for “Separate” or “Split Clip into Stems”.
3

Choose Song Stems or Drum Elements

Choose Song Stems or Drum Elements.
Stem separation only works on audio clips. If a MIDI clip is selected, the command palette will show a hint to select an audio clip instead.

Output

Gridshift creates a group track with one child track per stem, time-aligned with the original clip. The original clip remains unchanged. Each stem track is a regular sample track — you can apply effects, time-stretch, edit in Focus Mode, or use as a sidechain source like any other track.

In-Place Stems in the Sample Editor

If you want to mix the stems as a single instrument rather than arrange them independently, generate them inside the Sample Editor instead. The Sample Editor’s Stems mode bakes the same models against the current sample and exposes per-stem volume, mute, and solo controls on a stack of lanes — all on the original track.
Use this flowWhen
Split Clip into Stems (this page)You want each stem on its own track for independent arrangement, effects, and routing
Stems mode in the Sample EditorYou want to rebalance the parts inside one sample (e.g. lower the vocals on a loop) without adding tracks
Last modified on June 7, 2026