Custom Voice

POST

/kling/v1/general/custom-voices

Create a custom voice for use in later speech synthesis and related features.

Authorizations

bearer

Type

HTTP (bearer)

Request Body

application/json

object

Voice name

URL to fetch the voice source file.
Supports .mp3, .wav, .mp4, and .mov audio/video files.
The recording must contain a single clean human voice with no background noise, between 5 and 30 seconds long.

Historical work ID; reference a past generation to supply audio material.

Only videos that meet all of the following may be used for custom voice training:

• Generated with model version V2.6 and sound set to on
• Created via the digital human API
• Created via the lip-sync API

The audio must contain a single clean human voice with no background noise, between 5 and 30 seconds long.

Callback URL when the task completes.

Caller-defined task ID for reconciliation and lookup.

Responses

OK

Content-Type

application/json

POST

/kling/v1/general/custom-voices

Playground

Authorization

bearer

Body

Samples