StemSplit icon

StemSplit

Separate audio tracks into individual stems (vocals, drums, bass, etc.) using StemSplit AI

Actions5

Separate Stems (Wait for Completion)

AI-generated

Summary

Separate audio into stems and wait for completion, returning download URLs.

Inputs

  • Input Source (required) — Choose whether the audio is provided as a binary file from a previous node or via a publicly accessible URL.
  • Input Binary Property (required) — Name of the binary property containing the audio file (used when Input Source is binary).
  • Audio URL (required) — Publicly accessible URL of the audio file to separate (used when Input Source is URL).
  • Output Type (required) — Which stems to separate: Vocals + Instrumental, Vocals Only, Instrumental Only, Four Stems, or Six Stems.
  • Quality (required) — Processing quality: Best, Balanced, or Fast (Best required for Six Stems).
  • Output Format (required) — Audio format for output stem files: MP3, WAV, or FLAC.
  • Additional Options.File Name — Optional override for the file name used for the job.
  • Additional Options.Metadata — Optional custom metadata object echoed back in job responses.
  • Timeout (Seconds) (required) — Maximum number of seconds to wait for the job to complete before failing.
  • Poll Interval (Seconds) (required) — How often (in seconds) to check the job status.
  • Credentials (required) — StemSplit API credentials (API key) for authentication.

Output shape

object

For each input item the node returns an object containing job metadata plus download URLs for each requested stem. Stem URLs are named after the stem (e.g., vocalsUrl, drumsUrl) with a matching *ExpiresAt field indicating when the URL expires. The exact set of stem fields depends on the selected Output Type (BOTH, VOCALS, INSTRUMENTAL, FOUR_STEMS, SIX_STEMS).

Examples

Example 1: Separate a song into vocals and instrumental using a binary file from a previous node.

Input Source: Binary File; Input Binary Property: data; Output Type: BOTH; Quality: BEST; Output Format: MP3; Timeout: 600s; Poll Interval: 5s.

Example 2: Extract only the vocal track from a publicly available URL.

Input Source: URL; Audio URL: https://example.com/song.mp3; Output Type: VOCALS; Quality: Balanced; Output Format: WAV; Timeout: 300s; Poll Interval: 10s.

Links

Discussion