Music for a Short cut from your own video: Audio Library is off

You cannot add music or other sounds from the Audio Library to a Short made through Remix > Edit into a Short. YouTube's help page says so directly. If you want a music bed under a Short cut from your own video, build the Short as a file, with a soundtrack in the mix, and upload it as a normal Short.

The YouTube rules below come from the help page for creating Shorts from your videos, read 2026-10-02. The Sume side comes from the Music 1.0 and Timeline 1.0 docs.

Why is there no music in Edit into a Short?

The page lists it under key restrictions: "You cannot use music or other sounds from our Audio Library" on these Shorts, unlike standard Shorts creation. It gives no reason, and I will not guess one.

The page also does not say whether audio you brought in some other way is allowed in that flow. Treat that as unknown, and test in the app before you build a batch around it.

What are my options for sound?

There are three, and they differ in what you give up.

How do I generate a music bed with Sume?

Music 1.0 generates music from a text prompt, with no duration parameter; you steer length in the prompt, for example by asking for a 45-second track. The docs advise ending the brief with "Instrumental, no vocals." when you want a bed under speech. New integrations are pointed at POST /v1/music-router/generate.

Whether a generated track is right for your channel is your call. If you use AI music, read YouTube AI music detection and disclosure first.

How do I mix the music under the voice in Timeline?

Timeline 1.0 takes an optional soundtrack with url, gain_db, loop, fade_out_seconds (up to 10) and duck_db (0-20). Ducking needs a real audio spine, not silence, so the voice or the clip's own audio goes in audio.url. Both files must be on media.sume.com.

A 30-second Short with a looped bed 18 dB down that ducks 12 dB under the voice:

curl -X POST https://api.sume.com/v1/timeline-1.0/render \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: short-with-bed-001" \
  -d '{
    "audio": {"url": "https://media.sume.com/artifacts/artf_demo/voice.wav", "duration_seconds": 30},
    "video": [{"source_url": "https://media.sume.com/artifacts/artf_demo/cut.mp4", "start": 0, "duration": 30}],
    "soundtrack": {"url": "https://media.sume.com/artifacts/artf_demo/bed.mp3", "gain_db": -18, "loop": true, "fade_out_seconds": 2, "duck_db": 12}
  }'

What do I lose by leaving the in-app flow?

You give up the in-app text and filter tools and, as far as the page I read says, the automatic link back to the original video. You gain control over the mix and a file you keep. Sume does not upload to YouTube, so you upload the MP4 yourself.

If the music is purely for atmosphere, a standard Shorts upload with Audio Library sound is often simpler; use the file route when you need ducking, a fixed length or a repeatable batch.

A short checklist for a music bed

Write the music brief with a number for tempo and a named mood, and end with the instrumental instruction. Generate, listen, and reject tracks that fight the voice. Set the bed 15 to 20 dB below the voice with gain_db, add duck_db so it dips when someone talks, and give it a fade_out_seconds so it does not stop dead at the cut.

Run POST /v1/timeline-1.0/plan before the render. It is unbilled and returns duration_seconds and billable_minutes, so you know the cost of the render before you reserve it. The Timeline docs put the public rate at $0.10 per output minute.