Lyria 3.5 outputs MP3 or WAV: what you get from Sume music

Google's page lists MP3 by default or WAV for Lyria 3.5 at "44.1 kHz high-fidelity stereo". The Sume docs list no format field on the music request: the result's audio artifact is typically audio/mpeg on media.sume.com. A wav file may come from a one-part concat in timeline audio, whose default output is wav; the docs do not list accepted input formats, so test one job.

Sume details are from the Music 1.0 docs and Timeline audio docs, read 2026-09-30.

What does each side list?

Format choices differ at the request, not the sample rate claim.

How do I get a wav from a Sume track?

Timeline audio takes parts[] of one to twenty items, so a single part is a valid concat. Send the music artifact URL as the only part. The docs require this workspace's media.sume.com audio and do not say whether a generated music artifact is accepted directly or must be imported first (POST /v1/media-imports), so try one job before building on it. The result is kind: timeline_audio with an audio_url.

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: music-to-wav-001" \
  -d '{
    "operation": "concat",
    "parts": [
      { "url": "https://media.sume.com/artifacts/artf_demo/track.mp3" }
    ],
    "output": { "format": "wav" }
  }'

Does a wav restore detail?

No. Re-encoding a lossy file into wav gives a sample-exact container, not the detail that was dropped. The reason to do it is joining or lip-sync work: the docs say to keep wav when the file will be joined again, and that mp3 output re-adds priming padding at every edge. The job costs $0.01 flat.

How do I submit the job and fetch the track?

A music request takes mode async, sync, subscribe or webhook. With sync or subscribe, wait_timeout_seconds is 0 to 30; with webhook, webhook_url must be a public HTTPS callback. Send an Idempotency-Key on the submit, then poll GET /v1/jobs/{job_id}/status and read GET /v1/jobs/{job_id}/result. The audio is the entry in result.artifacts[] where type is audio, hosted on media.sume.com; raw provider URLs are not public outputs. Typically the file is audio/mpeg, per the Music 1.0 docs.

The prompt is 1 to 5000 characters. Put exclusions in the positive prompt ("Instrumental, no vocals"), because a non-empty negative_prompt is refused. Full field list: Music Router docs.