🎵 StemFlow User Guide

How to use StemFlow

StemFlow separates a music track into four AI-generated stems: vocals, drums, bass, and other instruments. Use it for remixing, practice, production, karaoke preparation, and music analysis.

🎤

Vocals

Isolated lead and backing vocal content.

🥁

Drums

Separated percussion and rhythmic elements.

🎸

Bass

Low-frequency bass content in a dedicated stem.

🎹

Other

Remaining instruments and non-bass accompaniment.

Quick start

Follow these steps to upload a track, run separation, and download the results.

Upload your audio

Select a WAV, MP3, M4A, FLAC, or OGG file from your device.

Start the split

Click Separate into Stems to send the file for AI processing.

Wait for processing

Small files finish quickly. Longer tracks may take several minutes.

Download the output

Download each WAV stem individually or use the ZIP package for all four stems together.

What happens behind the scenes

StemFlow uses an AI source-separation pipeline to analyse the full track and package the output for download.

Upload

Your audio file is uploaded from the website to the StemFlow processing service.

AI separation

The AI model isolates vocals, drums, bass, and other instruments from the full song.

Packaging

The separated stems are saved and prepared as WAV downloads and as a ZIP archive.

Delivery

When processing is complete, the site displays stem cards with direct download options.

Expected processing time

Processing time depends mainly on audio duration and file size. The table below reflects recent test results.

Size band	Example file	Typical processing time	Notes
Small	About 10–15 MB About 1–2 minutes audio	About 30–60 seconds	Usually completes quickly.
Medium	About 20–40 MB About 3–5 minutes audio	About 2–3 minutes	Allow time for upload and full separation.
Large	About 60 MB About 8 minutes audio	About 5–7 minutes	Recent test: a 62 MB, 8-minute file completed in about 6 minutes 55 seconds.

⚠️

Important warnings for large files

Large files can take several minutes. Do not refresh the page or close the browser tab during processing.
Files above the upload limit may be rejected. If this happens, trim the track or export a smaller version before retrying.
Longer tracks naturally take longer to process because the full song is analysed.
If the process appears slow, it usually means separation is still running. This is expected for longer audio.

Download options

Choose the format that best fits your workflow.

⬇️

Download WAV

Use the download button next to each stem card to save one isolated stem as a WAV file.

🗂️

Download All Stems (ZIP)

Download a ZIP package containing vocals, drums, bass, and other stems in one file.

Best practices

Use the cleanest source file you have. Higher-quality audio usually gives better separation.
For long songs, allow extra time before assuming the process has stalled.
If a file is rejected for size, convert it to a smaller format or trim the section you need.
The ZIP package is the fastest option for full DAW import workflows.

Troubleshooting

Upload rejected: the file may be above the upload limit.
Processing is slow: this is normal for medium and large files.
Download does not start: retry the individual WAV button or re-run the split and use ZIP.
Unexpected interruption: reopen StemFlow and upload the track again.