Music prompt limits: 1-5000 characters and no duration field

The music prompt must be 1 to 5000 characters, and there is no duration parameter: Sume rejects duration and duration_seconds. You set length in the prompt itself, for example "a 2-minute track" or a section map such as [0:00-0:30] Intro: .... That is the whole contract, from Music Router.

Fields you can send

A timed prompt

The Music 1.0 page suggests writing the arc into the prompt: emotion, genre, tempo as a number, key, two to four instruments, one named moment, and an era. The model treats these as creative direction, not guaranteed values, so listen to the result.

curl -X POST https://api.sume.com/v1/music-router/generate \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: timed-prompt-001" \
  -d '{
    "prompt": "Warm lo-fi, 84 BPM, C minor. Dusty Rhodes, brushed drums. [0:00-0:10] Intro: Rhodes alone. [0:10-0:30] Groove: drums and a muted trumpet answer. A 30-second track. Instrumental, no vocals."
  }'

What comes back

The job returns an audio artifact in result.artifacts[] where type is audio, usually audio/mpeg on media.sume.com, and result.lyrics can carry the model-reported lyrics or section map. That map is metadata, not a measurement, so check the actual length of the file instead of trusting the timestamps you wrote. Music 1.0 is described as making structured songs of up to a few minutes, so a "2-minute track" is inside the stated range.

Budgeting the characters

Five thousand characters is long enough for a full section map; most prompts need a tenth of that. The price does not depend on prompt length, so a longer, more specific brief costs the same as a short one.

Each accepted generation is $0.125. If the first take misses, rewrite the flagged part of the prompt and keep the musical brief; a generic fallback usually gives a worse bed than a second try.