ElevenLabs Music Generation

Generate music from text prompts - supports instrumental tracks, songs with lyrics, and fine-grained control via composition plans.

Setup: See Installation Guide. For JavaScript, use @elevenlabs/* packages only.

All examples below default to music_v2, the current generation model. Pass model_id="music_v1" only when explicitly requested to.

Workspace BGM Workflow

For video background music in this workspace:

1. Read Generated Music Registry before calling the API. Search by use case, mood, BPM, tags, and prompt. Audition and reuse a suitable recorded audio file when available. Reusing a prompt alone still triggers a paid generation; reuse its linked audio file to avoid spending credits.
2. Generate only when the registry has no suitable track. Keep new BGM requests at or below 30 seconds unless the user explicitly requests otherwise. Prefer 10–20 seconds for reusable loops.
3. For narration BGM, use force_instrumental=true and prompt for stable density, restrained melody, no intro or ending, no fade, no build or breakdown, and an ending that connects to the beginning. Specify BPM and a whole-bar duration when practical.
4. Preserve the exact API response as an immutable *-raw.* file. Never overwrite, trim, normalize, or re-encode it. Create loop masters, mix versions, and previews as separate derived files. Prefer a lossless 48 kHz WAV loop master.
5. Fix an otherwise suitable seam locally with a short circular crossfade instead of paying for another generation. Make a repeated preview to verify the seam before approving the loop.
6. After every successful generation, append an entry to Generated Music Registry before handoff. Record the exact prompt verbatim, date, project/use case, model, request duration, output format, instrumental setting, raw and derived file paths, measured durations, SHA-256 hashes, tags, approval status, and processing notes. Use workspace-relative paths and never record API keys.