YouTube Shorts splits audio and video: pull the track for $0.01

HeyOrca's roundup of YouTube news (read 2026-10-07) reports that the Shorts editor now separates the audio and video tracks. That helps when you edit inside YouTube. It does not help when the file is already rendered by an AI video model, an editor, or a colleague, and you need only its sound as a separate file. On Sume that is audio detach: one workspace video in, one new audio file out, the video untouched, a flat $0.01 per job.

The audio can then go back through timeline audio to be joined or sliced, or be used as the audio.url of a new render.

The call

The video must already be on media.sume.com in your workspace; Sume does not fetch from the open internet, so import it first with POST /v1/media-imports. An Idempotency-Key is required. Defaults are sample-exact WAV (pcm_s16le) and the source channel layout.

Check that the clip has sound before you pay: a source with no audio track fails with detach_source_has_no_audio, and a frames-free video inspect shows probe.has_audio.

curl -X POST https://api.sume.com/v1/audio-detach \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: short-audio-001" \
  -d '{"video_url": "https://media.sume.com/artifacts/artf_demo/short.mp4", "format": "wav"}'

Limits and prices, read 2026-10-07

A typical rework

Detach the sound ($0.01), split the first 3 seconds if you want to keep a hook line ($0.01), generate a new bed or voice, concat the parts into one file ($0.01), and render the Short again with that file as the soundtrack ($0.10 for a clip under a minute). The sound work costs three cents; the render is the larger line. Keep WAV between steps: the docs say MP3 adds encoder padding at every edge, which matters when you join files again.