POST
/
v1
/
sfx
/
generate
curl --request POST \
  --url https://os.gan.ai/v1/sfx/generate \
  --header 'Content-Type: application/json' \
  --header 'ganos-api-key: <api-key>' \
  --data '{
  "prompt": "<string>",
  "num_variations": 3,
  "duration_seconds": 1,
  "creativity": 0
}'
[
  {
    "sfx_inference_history_object": {
      "inference_id": "3c90c3cc-0d44-4b50-8888-8dd25736052a",
      "sfx_prompt": "<string>",
      "generation_params": {
        "duration_seconds": 123,
        "creativity": 123,
        "seed": 123
      },
      "created_at": "2023-11-07T05:31:56Z"
    },
    "wav_base64": "<string>"
  }
]

Authorizations

​
ganos-api-key
string
header
required

Use the API key generated from our website to programmatically interact with our endpoints. This key authorizes your requests, allowing you to access and utilize our services securely.

Body

application/json
​
prompt
string
required

A descriptive text input that outlines the desired characteristics of the sound effect to be generated. The minimum length of 5 characters and the maximum length is 256 characters should be provided.

Required string length: 5 - 256
​
num_variations
integer
default:3

Specifies the number of different variations of the sound effect to generate (1 to 4).

Required range: 1 <= x <= 4
​
duration_seconds
number
default:1

Defines the length of the sound effect in seconds. This ensures that the generated audio meets your specific timing requirements (1 to 20 seconds).

Required range: 1 <= x <= 20
​
creativity
number
default:0

Indicates the level of creative deviation from the prompt. A value of 0 means strict adherence to the prompt, while a value of 1 allows for maximum creative interpretation, resulting in more unique sound effects.

Response

200
application/json
Successful Response
​
sfx_inference_history_object
object
required

Includes the inference id which is the unique identifier for generated sound effects inference, audio prompt and the generation parameters provided as the input.

​
wav_base64
string
required

Provides audio as a Base64-encoded WAV file for easy text-based integration.