AI podcast intro music generator: one prompt, fixed price

Describe the whole intro in one prompt field, up to 5000 characters, and send it to the Music Router. Each generation is charged the fixed Music price ($0.125 per audio on the rate card), no matter which routable model runs it, so a long, detailed prompt costs the same as a short one.

What can I write in the prompt?

The Music Router docs (read 2026-10-01) say prompt is required and 1 to 5000 characters. Put exclusions in the positive prompt, since a non-empty negative_prompt is unsupported. A podcast intro usually needs a mood, instruments, tempo and a length hint, for example: "Upbeat acoustic bumper, claps and a bright guitar riff, 100 BPM, instrumental, no vocals, about 15 seconds."

Which model ids can I route to?

The public id is sume/music-router. Routable ids are sume/music-auto (the default when model is omitted), lyria-3.5 and lyria-3-pro. The job's request.routed_model tells you which engine ran.

How do I call it?

An Idempotency-Key header is used on the create call.

const res = await fetch("https://api.sume.com/v1/music-router/generate", {
  method: "POST",
  headers: {
    Authorization: "Bearer " + process.env.SUME_API_KEY,
    "Content-Type": "application/json",
    "Idempotency-Key": "podcast-intro-001",
  },
  body: JSON.stringify({
    model: "sume/music-auto",
    prompt:
      "Upbeat acoustic bumper, claps and a bright guitar riff, 100 BPM. About 15 seconds. Instrumental, no vocals.",
  }),
});
console.log(res.status, await res.json());

How do I fit the intro under my episode?

Once you have the file, a Timeline 1.0 render can lay it as a soundtrack bed with gain_db and a fade_out_seconds of up to 10, so the intro ends cleanly. Related: jingle generator for a podcast intro.