Read the Sume Music Router catalog: auto, Lyria 3.5 and Lyria 3 Pro

GET https://api.sume.com/v1/music-router/models returns three rows in data.models: sume/music-auto, lyria-3.5 and lyria-3-pro. Each has capability flags, a maximum prompt length and a pricing object with the provider list price and the billable price in micro-dollars per audio.

The Sume docs say every row is charged the one fixed Music price, so the catalog is mostly for checking which engine sume/music-auto resolves to.

A script that prints it

Run with Node 18 or later and SUME_API_KEY set.

async function main() {
  const res = await fetch("https://api.sume.com/v1/music-router/models", {
    headers: { "x-api-key": process.env.SUME_API_KEY ?? "" },
  });
  if (!res.ok) throw new Error("HTTP " + res.status);
  const { data } = await res.json();
  for (const m of data.models) {
    const p = m.pricing;
    console.log(
      m.id.padEnd(16),
      m.routing.padEnd(13),
      "resolves_to", m.resolves_to ?? "-",
      "max prompt", m.capabilities.max_prompt_characters,
      "billable $" + (p.billable_usd_micros_per_audio / 1e6).toFixed(3),
    );
  }
}
main().catch((e) => { console.error(e.message); process.exit(1); });

What to expect

From the repo catalog and docs, the rows are as below. The live response is the source of truth.

A catalog quirk worth reading twice

The Pro row lists a lower provider price than Lyria 3.5, yet the docs state that every Music Router model is charged one fixed Music price. Do not assume Pro is cheaper for you. Check a finished job's cost, or read billable_usd_micros_per_audio from your own response.

Generation itself takes a prompt of 1 to 5,000 characters, an optional image_url, and has no duration field. The output is an mp3 on media.sume.com.

Handling errors

A 401 means the key is missing, malformed or revoked. Send the key as x-api-key or as an Authorization Bearer header. Every error response carries an x-sume-request-id header that starts with req_, so log it and quote it if you contact support.

Use auto unless you are testing

sume/music-auto follows whatever Sume resolves, so your prompt keeps working if the default engine changes. Pin lyria-3.5 when you are comparing runs and need the engine fixed, as in a blind test of prompts. Because each try is a flat charge, it is cheap to run the same prompt on both explicit ids and listen.