Music and sound effects API

Music and sound effects use POST /generations with mode: "audio". Discover current audio models with GET /models?mode=audio and inspect their capabilities.

Music example

curl --fail-with-body https://app.verbatik.com/api/v1/generations \
  -H "Authorization: Bearer $VERBATIK_API_KEY" \
  -H 'Content-Type: application/json' \
  -H 'Idempotency-Key: campaign-music-001' \
  -d '{
    "mode":"audio",
    "model":"fal-ai/minimax-music/v2",
    "prompt":"An instrumental product-film soundtrack with warm piano and soft percussion",
    "delivery":"signed_url"
  }'

When the selected model supports lyrics or duration, provide the corresponding parameters values from its capabilities. Do not copy the old num_songs, tags-combination rules, or output-format fields into this API.

Sound effect example body

{
  "mode": "audio",
  "model": "fal-ai/elevenlabs/sound-effects/v2",
  "prompt": "Gentle rain on a canvas tent with distant wind",
  "parameters": { "duration": "5", "loop": true },
  "delivery": "signed_url"
}

Check the model catalog before choosing duration and looping settings. These requests require generations:write. Creation returns 202; poll GET /generations/{id} with generations:read for outputs.

Both music and sound effects consume workspace credits. Costs vary by model and settings; they do not use the legacy API's per-minute dollar wallet.