AI music reference artist song blocked: write the brief without one

If an AI music tool blocks a download because the track references another artist's song, the fix is to rewrite the prompt so it describes a sound instead of pointing at a song. ElevenLabs says Music v2.5 blocks downloads on tracks that reference other artists' songs. Sume's music brief avoids the problem by design: it describes emotion, genre, tempo, key, instruments, arc and era, with no artist name anywhere.

The ElevenLabs fact is from its Music v2.5 post, read 2026-09-29. The brief is from Music 1.0 and the Music Router docs. Sume's docs do not describe a download block, and this post does not guess what any service's filter matches.

What does a brief without an artist name look like?

Sume's Music docs give a seven-axis brief. Each axis replaces a reason people name an artist: the mood, the lineage, the pace, the palette, the shape.

What does a full prompt look like?

This is the example from the Music docs. It names no artist, and it ends with the instrumental clause the docs recommend.

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: no-artist-brief-001" \
  -d '{
    "prompt": "Hushed and slightly melancholic neo-soul nocturne, 72 BPM, D minor. Rhodes through tape wow, soft sub bass, brushed snare with rimshots, a single muted trumpet line. Late-night, dry and close, 1998 production. Instrumental, no vocals."
  }'

What if a request is rejected?

Sume's docs say that on a policy rejection you should revise the flagged content while keeping the musical brief, and retry only within your authorized budget. They also say not to strip the request down to a generic bed. So keep the tempo, key and instruments, and change only the part that was flagged.

Can I keep style words but drop the name?

Yes. The docs list genre and lineage terms such as neo-soul, drill, bossa nova, gugak fusion, synthwave and chamber folk as examples of what to write. Exclusions go in the positive prompt too: the Music docs suggest “no vocals, no spoken word” and reject a non-empty negative_prompt with HTTP 400.

How specific should the instruments and moments be?

Specific enough that the model has less reason to reach for a familiar reference. The docs ask for two to four instruments with texture, such as “Rhodes through tape wow” or “gayageum plucks”, and for one named moment in the arc, such as “breakdown to bass and claps at 0:20, full return at 0:28”.

The docs call these axes creative directions, not guaranteed output settings, and tell you to verify the generated audio. So listen to the result, and adjust the axes rather than adding a name.