Join two generated tracks without a gap: timeline audio concat

How do you join two AI-generated music tracks into one file? Use POST /v1/timeline-1.0/audio with operation: "concat" and a parts[] list. The Timeline audio docs say the join is sample-domain, so there is no silence at the seams and no re-synthesis, and the result is a durable media.sume.com file with the offsets for each part.

Parts are ordered, 1 to 20 of them, and every URL must already be audio in your workspace on media.sume.com. Import files first with POST /v1/media-imports.

The request

Each part is { url, source_in?, duration? }. Do not send url or ranges at the top level; that returns audio_concat_takes_no_url or audio_concat_takes_no_ranges. Idempotency-Key is required.

curl -X POST https://api.sume.com/v1/timeline-1.0/audio \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: concat-tracks-001" \
  -d '{
    "operation": "concat",
    "parts": [
      { "url": "https://media.sume.com/artifacts/artf_demo/track1.mp3" },
      { "url": "https://media.sume.com/artifacts/artf_demo/track2.mp3", "source_in": 0.1, "duration": 28 }
    ]
  }'

What comes back

A kind: timeline_audio result with one audio_url, a duration_seconds and segments[] holding index, start and duration_seconds. Those offsets are what you re-base video[].start against on a Timeline 1.0 render.

wav or mp3

The default is sample-exact pcm_s16le wav. mp3 is smaller but re-adds priming padding at every edge, so keep wav for any file you will join again. Beds from the router come back as audio/mpeg, so the first join of two of them is already an mp3 input; decide the output format with that in mind.

When not to mint a file

If the join is only needed inside one render, put the parts on audio.parts[] of the Timeline 1.0 request and skip this job. Mint a file only when you want to reuse it.