AI Sound Separation REST API Reference Documentation (1.1.0)

The AI Sound Separation REST API supports an algorithm for separating mixed source audio files into component tracks. It produces tracks designated as drums, bass, vocals, and "other".

Authentication

ApiKeyAuth

TechHub API uses an authentication token. Please use the TH_API_TOKEN in the header field. The token can be generated from the API Console developer page.

Security Scheme Type	API Key
Header parameter name:	X-TH_API-Key

Inference

Directly separate short audio

Separate audio into 4 tracks

This API is designed for audio file of small size.

Authorizations:

ApiKeyAuth (YOUR_TOKEN)

Request Body schema: application/json

Parameter 'audio' contains one or more audio files. Each audio file shall be encoded in base64 format. Parameter 'algorithm' contains algorithm option.

	Array of objects
	object

Responses

Request samples

Payload

Content type

application/json


{"audio": 
[
{"name": "10sec_audio.mp3",
"data": "data:audio/mpeg:base64,SUQzBAAAAAAgWlRFTkMAAAAMAAAATG...RhKaqqqqqqqqqqqqqqqqo="
}
],
"algorithm": 
{"name": "default",
"option": 
[
{"name": "targets",
"data": 
["vocals",
"drums",
"bass",
"other"
]
},
{"name": "start_point",
"data": 0
}
]
}
}

Response samples

200
400
413
500

Content type

application/json


{"status": 0,
"message": "string",
"output": 
[
{"name": "string",
"data": "string"
}
]
}