Lyria 3.5 takes 10 images: Sume music takes one image_url

Google's Gemini API page says Lyria 3.5 accepts "up to 10 images" with text, but Sume's music request takes one optional image_url, a public HTTPS image, or null to clear it. Choose the single still that matches the mood you want, or put the rest of the scene in words in the prompt.

Sume fields are from the Music 1.0 docs and Music Router docs, read 2026-09-30.

How do the two inputs compare?

Only the image count differs in this table; both accept text.

Which single image should I send?

The Music docs suggest passing the accepted scene still as image_url when it is appropriate, and preserving continuity when one consistent score is requested. For a moodboard, choose the frame that carries the mood and describe the others in the brief: emotion, genre, tempo, key, instruments, arc.

What does an image-conditioned request look like?

The Music docs show this shape. Sume treats the image as conditioning; it does not promise the music matches every detail.

curl -X POST https://api.sume.com/v1/music-1.0/generate \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: music-image-001" \
  -d '{
    "prompt": "Cinematic ambient underscore matching the mood of the reference still, instrumental only",
    "image_url": "https://example.com/moodboard.png"
  }'

Does that route through the Music Router?

Yes. The Music 1.0 page says every request now resolves through the Music Router, and job.model stays sume/music-1.0 on that route. New integrations should call POST /v1/music-router/generate, which accepts the same body.

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. The image is conditioning, not a guarantee of match; verify the audio.

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.