Audio Sound Effects

Generate non-speech audio from a text prompt. Synchronous, flat per-generation pricing.

POST /v1/audio/sfx

Audio Sound Effects

cURL

curl --request POST \
  --url https://kymaapi.com/v1/audio/sfx \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "model": "<string>",
  "prompt": "<string>",
  "duration_seconds": 123,
  "prompt_influence": 123,
  "response_format": "<string>"
}
'

Synchronous endpoint. Describe a sound — door slam, rain, sword unsheathing, ambient room — get back a clip. Pairs naturally with /v1/audio/music for video and game soundtracking.

curl -X POST https://kymaapi.com/v1/audio/sfx \
  -H "Authorization: Bearer $KYMA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "heavy wooden door slamming shut in stone hallway, with reverb",
    "duration_seconds": 3
  }' \
  --output door.mp3

Request

application/json body.

model · string · default: "elevenlabs-sfx"

SFX model. Currently only elevenlabs-sfx is shipped.

prompt · string · required

Description of the sound. Max 500 characters. Also accepts text as an alias.

duration_seconds · number

Optional duration 0.5 to 22 seconds. Omit to let the model auto-pick a sensible length.

prompt_influence · number · default: "0.3"

How literal the prompt should be (0 = creative, 1 = strict). Range 0–1.

response_format · string · default: "mp3_44100_128"

Audio format. Same options as audio/speech.

save_to_blob · string

Optional query param. Set to 1 to have Kyma upload the resulting MP3 to Vercel Blob, write a multimodal_jobs row, and return JSON { job_id, kind: "audio", url, duration_sec, cost_usd, balance_usd } instead of streaming bytes. Used by Canvas and Muse audio kinds — most direct API callers don’t need this.

Response

Default (streaming)

200 OK with audio bytes. Flat pricing — duration doesn’t change the cost.

Header	What
X-Kyma-Model	elevenlabs-sfx
X-Kyma-Cost-USD	flat $0.027 per generation
X-Kyma-Balance-USD	remaining balance

With ?save_to_blob=1

200 OK with JSON:

{
  "object": "audio.generation",
  "job_id": "mmj_...",
  "kind": "audio",
  "model": "elevenlabs-sfx",
  "url": "https://blob.vercel-storage.com/mmj/.../mmj_....mp3",
  "duration_sec": 3,
  "cost_usd": 0.027,
  "balance_usd": 47.97
}

Errors

Status	error.code	When
400	not_an_sfx_model	model is not an SFX SKU
400	prompt_too_long	prompt > 500 chars
400	invalid_duration	duration_seconds outside [0.5, 22]
401	auth_error	missing or invalid API key
402	billing_error	balance too low
502	provider_error	upstream provider failure

See also

• POST /v1/audio/music — full music tracks
• POST /v1/audio/speech — voice narration