On-device vocal and instrumental separation

One recording.Two useful tracks.

Import a video from Photos or a local audio file. Separate vocals and accompaniment on your iPhone or iPad, preview the result, adjust each level, and export without uploading the source file.

  • On-device
  • Photo-library video
  • Fast / Fine modes
  • Export results
Feature illustration
9:41
ShengXi
Source audio03:28 · local file
Vocals
72%
Instrumental
84%
Source stays on your device
Separate audio
A focused workflow

From import to export, without unnecessary steps

ShengXi focuses on two-track separation. Choose a source and processing mode, then preview vocals and accompaniment independently. Interface graphics on this page are feature illustrations, not pixel-perfect screenshots.

01

Start with video or audio

Choose a video from Photos or a local audio file. Audio extraction and separation continue on your device.

  • Photo-library video
  • Local audio file
  • Recent result history
02

Fast and Fine modes

Use Fast mode for a quick check, then Fine mode when the source needs a more careful pass. Time and quality vary by source.

  • Quick preview
  • Fine processing
  • Honest quality limits
03

Preview each result

Listen to the mix or adjust vocal and instrumental levels independently, then save, share, or remove recent results.

  • Mixed preview
  • Independent levels
  • Save and share
Three simple steps

Spend more time practicing and creating

01

Choose a source

Select a video or audio file you have permission to process.

02

Separate on device

Pick Fast or Fine mode. Your source does not need to be uploaded.

03

Preview and export

Listen to each track, adjust levels, and save or share the result.

Built for real tasks

A simpler step in your audio workflow

A singer practicing at home with a phone and headphones

Covers and singing practice

Create a practice backing track from music you are permitted to use.

Music study

Hear vocals and accompaniment separately to study melody, rhythm, and arrangement.

Short-video audio

Prepare separate voice or background tracks from footage you own or are licensed to edit.

Private local work

Process recordings and unreleased material without sending the source to a server.

A creator preparing short-video audio at a desk
Local-first processing

Your source does not need to leave your device

Audio extraction, model separation, and export are completed on device. You are still responsible for having permission to process and publish any music or video you use.

Solve one specific problem

Latest guides

Browse all guides →
The first practical guides are being prepared.
Solve one specific problem

Popular topics

Vocal removerBacking tracksRemove vocalsVideo audioOffline processingCover practice
Know before you process

Frequently asked questions

Will every song separate perfectly?

No. Mixing, compression, noise, and source quality can leave artifacts or change tone.

Is my source uploaded?

No. The current version extracts, separates, and exports audio on device.

What can I import?

Videos from Photos and local audio files. Exact selectable formats depend on iOS and the current app version.

Which mode should I use?

Try Fast mode first, then Fine mode when you want a more careful pass.

Feedback and support

Tell us the source type and device when something goes wrong

Do not email full songs you are not allowed to share. File type, device model, and OS version are usually enough to start troubleshooting.

developer@leleyes.com