Skip to main content
POST
/
v1
/
voice_clone
Voice Clone
curl --request POST \
  --url https://api.minimax.io/v1/voice_clone \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: <content-type>' \
  --data '
{
  "file_id": 123456789,
  "voice_id": "<voice_id>",
  "clone_prompt": {
    "prompt_audio": 987654321,
    "prompt_text": "This voice sounds natural and pleasant."
  },
  "text": "A gentle breeze sweeps across the soft grass(breath), carrying the fresh scent along with the songs of birds.",
  "model": "speech-2.8-hd",
  "text_validation": "This voice sounds natural and pleasant.",
  "accuracy": 0.7,
  "need_noise_reduction": false,
  "need_volume_normalization": false,
  "aigc_watermark": false
}
'
{
  "input_sensitive": false,
  "input_sensitive_type": 0,
  "demo_audio": "",
  "extra_info": {
    "audio_length": 11124,
    "audio_sample_rate": 32000,
    "audio_size": 179926,
    "bitrate": 128000,
    "word_count": 18,
    "usage_characters": 18
  },
  "base_resp": {
    "status_code": 0,
    "status_msg": "success"
  }
}

Authorizations

​
Authorization
string
header
required

HTTP: Bearer Auth

Headers

​
Content-Type
enum<string>
default:application/json
required

The media type of the request body. Must be set to application/json to ensure the data is sent in JSON format.

Available options:
application/json

Body

application/json

Voice clone request parameters

​
file_id
integer<int64>
required

The file_id of the audio to be cloned, obtained through the File Upload API.

Uploaded files must comply with the following rules:

  • Accepted audio formats: mp3, m4a, wav
  • Audio duration: at least 10 seconds, no longer than 5 minutes
  • File size: no larger than 20 MB
  • If this parameter is used, both child attributes (prompt_audio, prompt_text) are required
​
voice_id
string
required

The voice_id of the cloned voice. Example: "MiniMax001". When defining a custom voice_id, note the following rules:

  • Length range: [8, 256]
  • Must start with an English letter
  • Can contain letters, digits, -, and _
  • Cannot end with - or _
  • Must not duplicate an existing voice_id, otherwise an error will occur
​
clone_prompt
object

Voice cloning parameters. Providing this field helps improve the similarity and stability of synthesized voice. If used, you must also upload a short sample audio clip (less than 8s, supported formats: mp3, m4a, wav) along with its corresponding transcript.

​
text
string

Optional preview text, up to 1000 characters. The cloned voice will be used to read the text, and an audio preview link will be returned. Note: Preview requests are charged based on character count, consistent with T2A pricing.

  • Interjection tags: Only supported when using speech-2.8-hd or speech-2.8-turbo models. Supported interjections: (laughs), (chuckle), (coughs), (clear-throat), (groans), (breath), (pant), (inhale), (exhale), (gasps), (sniffs), (sighs), (snorts), (burps), (lip-smacking), (humming), (hissing), (emm), (whistles), (sneezes), (crying), (applause).
​
model
enum<string>

Specifies which voice synthesis model to use for generating the preview audio. Required when the text field is provided.

Available options:
speech-2.8-hd,
speech-2.8-turbo,
speech-2.6-hd,
speech-2.6-turbo,
speech-02-hd,
speech-02-turbo,
speech-01-hd,
speech-01-turbo
​
language_boost
enum<string>

Controls whether recognition for specific minority languages and dialects is enhanced. Default is null. If the language type is unknown, set to "auto" and the model will automatically detect it.

Available options:
Chinese,
Chinese,Yue,
English,
Arabic,
Russian,
Spanish,
French,
Portuguese,
German,
Turkish,
Dutch,
Ukrainian,
Vietnamese,
Indonesian,
Japanese,
Italian,
Korean,
Thai,
Polish,
Romanian,
Greek,
Czech,
Finnish,
Hindi,
Bulgarian,
Danish,
Hebrew,
Malay,
Persian,
Slovak,
Swedish,
Croatian,
Filipino,
Hungarian,
Norwegian,
Slovenian,
Catalan,
Nynorsk,
Tamil,
Afrikaans,
auto
​
text_validation
string

Optional. The expected transcript of the cloning sample audio (matching the content of file_id or clone_prompt.prompt_audio). When provided, the audio is sent to ASR and the recognized text is compared against text_validation. If the similarity is below accuracy, the request is rejected with status code 1043 (The asr similarity check failed). Maximum length: 200 characters.

Maximum string length: 200
​
accuracy
number<double>
default:0.7

Optional. Similarity threshold used by the ASR validation triggered by text_validation. Valid range: [0, 1]. When omitted or set to 0, defaults to 0.7.

Required range: 0 <= x <= 1
​
need_noise_reduction
boolean
default:false

Indicates whether to enable noise reduction.

​
need_volume_normalization
boolean
default:false

Indicates whether to enable volume normalization.

​
aigc_watermark
boolean
default:false

Indicates whether to append an AIGC watermark tone to the end of the synthesized preview audio.

Response

200 - application/json

Successful response

​
input_sensitive
object

Content safety check result

​
demo_audio
string

If both text and model are provided, this field returns a URL to the preview audio. Otherwise, it will be empty.

​
extra_info
object

Preview audio metadata and billing info. Returned only when text is provided (i.e. preview synthesis happened and was billed). Field shape is aligned with /v1/t2a_v2.

​
base_resp
object