Description

General-purpose speech multitask engine.

Version

1.0

Media Requirements

Supported audio formats: WAV, M4A, MP4, MP3, AAC, and others.

API URL

Production environment: https://openapi.meitu.com
Task submission endpoint: https://openapi.meitu.com/api/v1/sdk/sync/push
Task name (`task`): /v1/ai_audio_engine/481991
Task type (`task_type`): formula

Method

POST

Content-Type: application/json

Authentication

Open Platform API Signing

Request Parameters

Required	Parameter	Type	Description
Required	task	String	/v1/ai_audio_engine/481991
Required	task_type	String	formula
Required	init_images	Object[]	Multimedia files
Required	params	String	Inference parameters
Optional	sync_timeout	Int	Default: `30`. Synchronous timeout. If the API times out, `status` is `9`; use the Task Query API to retrieve the result.

The structure of each multimedia item in init_images is described below.

Required	Parameter	Type	Description
Required	url	String	Audio URL or base64 data
Required	profile	object	Audio attribute information

The profile object is described below.

Required	Parameter	Type	Description
Required	media_profiles	object	Audio attribute information
Required	version	string	Fixed value: `v1`

The fields in media_profiles are described below.

Required	Parameter	Type	Description
Required	media_data_type	String	`url` indicates a URL; the corresponding audio type indicates base64 data.

params is a JSON string whose decoded structure is described below.

Required	Parameter	Type	Description
Optional	rsp_media_type	String	Default: `url`
Required	parameter	object	Inference parameters

The fields in parameter are described below.

Required	Parameter	Type	Description
Required	task_type	String	Task type. Supported values are listed below.

Supported task_type values:

Value	Description	Dependencies
lufs	Retrieve audio loudness	-
trans_audio	Transcribe and translate audio; return the translation and transcription	-
ducking	Vocal-ducking keyframes	-
train	Train and save a voice profile	ai-audio-timbre, ai-audio-spliter, ai-audio-asr
synth	Generate audio using an existing voice profile	ai-audio-timbre, ai-audio-spliter, ai-audio-asr
synth_once	Clone and generate once without saving the voice profile	ai-audio-timbre, ai-audio-spliter, ai-audio-asr
edit	Voice editing	ai-audio-timbre, ai-audio-spliter, ai-audio-asr
trans	Translate and generate using an existing voice profile	ai-audio-timbre, ai-audio-spliter, ai-audio-asr
trans_once	Extract a voice profile and translate audio in real time	ai-audio-timbre, ai-audio-spliter, ai-audio-asr
conversion	Voice conversion	ai-audio-spliter, ai-audio-conversion

Detailed parameters for each task_type:

1. lufsr (section name in the source document; the `task_type` value is `lufs`, which retrieves audio loudness)

Required	Parameter	Type	Description
No	app	String	Caller application
No	uid	String	`uid`; pass `0` when unavailable
No	gid	String	gid

2. trans_audio (speech transcription and translation)

Required	Parameter	Type	Description
Yes	from	String	Source language
Yes	to	String	Target language
Yes	client_id	Int	Application client ID
No	asr_from	String	ASR provider; default: `tencent`
No	trans_type	String	Translation provider; default: `tencent`
No	energy_threshold	Double	Audio RMS threshold; default: `0`, which skips task processing
No	app	String	Caller application
No	uid	String	`uid`; pass `0` when unavailable
No	gid	String	gid

3. ducking (vocal-ducking keyframes)

Required	Parameter	Type	Description
No	app	String	Caller application
No	uid	String	`uid`; pass `0` when unavailable
No	gid	String	gid

4. train (voice training and voice-profile storage)

Required	Parameter	Type	Description
No	prompt_lang	String	Language of the input audio. The default is `zh` in China and `en` outside China; pass `en` for English.
No	model_name	String	Supported values: `qwen`, `index_tts2`. The production model is `index_tts2`.
No	version	String	Model version. Currently supports `v1`, `v2`, and `v3`; default: `v1`.
No	train_limit	Float	Training segment length in seconds; default: `10s`. Set to `0` to extract features from the entire input audio.
No	spliter	Bool	`True`/`False`; whether the input requires denoising
No	spliter_type	String	Denoising model type
No	asr_text	JSON	If provided, `asr_text` is used by default.
No	app	String	Caller application
No	uid	String	`uid`; pass `0` when unavailable
No	gid	String	gid

5. synth (generation using an existing voice profile)

Required	Parameter	Type	Description
Yes	spk_id	String	`spk_id` produced by model training
Yes	text	String	Text to synthesize
No	sp_speed	Float	Speed multiplier for output-audio `sp`; default: `1.0`
No	speed	Float	Output-audio speed multiplier; default: `1.0`. To double the speed, pass `1/2 = 0.5`.
No	duration	Float	Desired output-audio duration; default: `0`, which preserves the duration generated by the model
No	use_ssml	Bool	`True`/`False`; whether to enable SSML; default: `True`
No	gen_lang	String	Output-audio language; supports `zh` and `en`; default: `zh`
No	gen_size	Int	Synthesize audio in batches; number of characters per batch
No	app	String	Caller application
No	uid	String	`uid`; pass `0` when unavailable
No	gid	String	gid

6. synth_once (one-time cloning and generation without saving the voice profile)

Required	Parameter	Type	Description
Yes	text	String	Text to synthesize
No	sp_speed	Float	Speed multiplier for output-audio `sp`; default: `1.0`
No	speed	Float	Output-audio speed multiplier; default: `1.0`
No	duration	Float	Desired output-audio duration; default: `0`, which preserves the duration generated by the model
No	use_ssml	Bool	`True`/`False`; whether to enable SSML; default: `True`
No	prompt_lang	String	Language of the input audio. The default is `zh` in China and `en` outside China; pass `en` for English.
No	gen_lang	String	Output-audio language; supports `zh` and `en`; default: `zh`
No	gen_size	Int	Synthesize audio in batches; number of characters per batch
No	model_name	String	Supported values: `qwen`, `index_tts2`. The production model is `index_tts2`.
No	version	String	Model version. Currently supports `v1`, `v2`, and `v3`; default: `v1`.
No	train_limit	Float	Training segment length in seconds; default: `10s`. Set to `0` to extract features from the entire input audio.
No	spliter	Bool	`True`/`False`; whether the input requires denoising
No	spliter_type	String	Denoising model type
No	asr_text	JSON	If provided, `asr_text` is used; otherwise the internal ASR service is invoked.
No	app	String	Caller application
No	uid	String	`uid`; pass `0` when unavailable
No	gid	String	gid
No	emo_control_method	int	Emotion-control method; default: `0`. `0`: emotion from reference audio; `1`: manual emotion vector; `2`: independent emotion reference audio. See below.
No	emo_vec	float[]	Eight-dimensional manual emotion vector; required when `emo_control_method=1`
No	emo_ref_path	String	Emotion-reference audio URL; required when `emo_control_method=2`
No	emo_weight	float	Emotion-reference audio weight; optional when `emo_control_method=2`; default: `1.0`; recommended range: `0.0–1.6`

Emotion Control

emo_control_method controls the emotion source for synthesized audio and supports the following three modes:

Value	Mode	Description	Additional parameters
0	Reference-audio emotion (default)	Extract emotion from the voice reference audio uploaded in `init_images`	None
1	Manual emotion vector	Precisely control emotion intensity with an eight-dimensional vector	emo_vec (required)
2	Independent emotion reference audio	Use another audio clip to control emotion while retaining the voice from the reference audio	emo_ref_path (required), emo_weight (optional)

Order of the eight dimensions in emo_vec: [happiness, anger, sadness, fear, disgust, depression, surprise, calmness]

Constraints:

emo_vec must be a numeric array of length 8.
The sum of the eight weights must not exceed 1.5.
If emo_control_method is omitted, it defaults to 0, meaning the emotion is extracted from the voice reference audio.

7. edit (voice editing)

Required	Parameter	Type	Description
Yes	text	String	Text to synthesize
Yes	start_time	Float	Start time of the audio segment to replace (milliseconds)
Yes	end_time	Float	End time of the audio segment to replace (milliseconds)
No	sp_speed	Float	Speed multiplier for output-audio `sp`; default: `1.0`
No	speed	Float	Output-audio speed multiplier; default: `1.0`
No	duration	Float	Desired output-audio duration; default: `0`, which preserves the duration generated by the model
No	use_ssml	Bool	`True`/`False`; whether to enable SSML; default: `True`
No	prompt_lang	String	Language of the input audio. The default is `zh` in China and `en` outside China; pass `en` for English.
No	gen_lang	String	Output-audio language; supports `zh` and `en`; default: `zh`
No	gen_size	Int	Synthesize audio in batches; number of characters per batch
No	model_name	String	Supported value: `voice_mirror`. The production model is `voice_mirror`.
No	version	String	Model version. Currently supports `v1`, `v2`, and `v3`; default: `v3`.
No	train_limit	Float	Training segment length in seconds; default: `10s`. Set to `0` to extract features from the entire input audio.
No	spliter	Bool	`True`/`False`; whether the input requires denoising
No	spliter_type	String	Denoising model type
No	asr_text	JSON	If provided, `asr_text` is used; otherwise the internal ASR service is invoked.
No	mix	Bool	`True`/`False`; whether to overlay background audio; default: `False`
No	single	Bool	`True`/`False`; whether to use single-speaker mode. If `True`, the entire audio is used to extract model features; if `False`, audio beginning at `start_time` is used.
No	app	String	Caller application
No	uid	String	`uid`; pass `0` when unavailable
No	gid	String	gid

8. trans (translated generation using an existing voice profile)

Required	Parameter	Type	Description
Yes	process_info	JSON Array	Synthesized-audio information containing `start_time`, `end_time`, `text`, and `trans_text`
Yes	spk_id	String	Existing voice-profile ID
Yes	from	String	Source language; default: Chinese (`zh`)
Yes	to	String	Target language; default: English (`en`)
No	use_ssml	Bool	`True`/`False`; whether to enable SSML; default: `True`
No	model_name	String	Model name
No	version	String	Model version
No	app	String	Caller application
No	uid	String	`uid`; pass `0` when unavailable
No	gid	String	gid

9. trans_once (real-time voice extraction and translation)

Required	Parameter	Type	Description
Yes	process_info	JSON Array	Synthesized-audio information containing `start_time`, `end_time`, `text`, and `trans_text`
No	from	String	Source language; default: Chinese (`zh`)
No	to	String	Target language; default: English (`en`)
No	use_ssml	Bool	`True`/`False`; whether to enable SSML; default: `True`
No	model_name	String	Model name
No	version	String	Model version
No	sp_speed	Float	Speed multiplier for output-audio `sp`; default: `1.0`
No	speed	Float	Output-audio speed multiplier; default: `1.0`
No	duration	Float	Desired output-audio duration; default: `0`, which preserves the duration generated by the model
No	train_limit	Float	Training segment length in seconds; default: `10s`. Set to `0` to extract features from the entire input audio.
No	spliter	Bool	`True`/`False`; whether the input requires denoising
No	spliter_type	String	Denoising model type
No	asr_text	JSON	If provided, `asr_text` is used; otherwise the internal ASR service is invoked.
No	single	Bool	`True`/`False`; whether to use single-speaker mode. If `True`, the entire audio is used to extract model features; if `False`, audio beginning at `start_time` is used.
No	app	String	Caller application
No	uid	String	`uid`; pass `0` when unavailable
No	gid	String	gid

10. conversion (voice conversion)

Required	Parameter	Type	Description
No	spk_id	String	Voice-profile ID (supported values: `ziwei`, `silang`, `wukong`, `baishe`)
No	suffix	String	Returned audio format; default: `mp3`. Pass `wav` to return WAV audio.
No	app	String	Caller application
No	uid	String	`uid`; pass `0` when unavailable
No	gid	String	gid

Request Examples

Voice cloning — default mode (use emotion from the reference audio)

{
	"task": "/v1/ai_audio_engine/481991",
	"task_type": "formula",
	"init_images": [{
		"url": "http://action-public.meitudata.com/voice_clone/68c12c8a20160928587d19j69h5423.mp3",
		"profile": {
			"media_profiles": {
				"media_data_type": "url"
			},
			"version": "v1"
		}
	}],
	"params": "{\"rsp_media_type\":\"url\",\"parameter\":{\"task_type\":\"synth_once\",\"text\":\"我是你克隆的专属声音，快来用我制作视频吧。This is the voice of your clone, please use it to create videos.\"}}",
	"sync_timeout": 30
}

Voice cloning — manual emotion vector (primarily sadness)

{
	"task": "/v1/ai_audio_engine/481991",
	"task_type": "formula",
	"init_images": [{
		"url": "http://action-public.meitudata.com/voice_clone/68c12c8a20160928587d19j69h5423.mp3",
		"profile": {
			"media_profiles": {
				"media_data_type": "url"
			},
			"version": "v1"
		}
	}],
	"params": "{\"rsp_media_type\":\"url\",\"parameter\":{\"task_type\":\"synth_once\",\"text\":\"今天天气真不错，你怎么知道的？\",\"emo_control_method\":1,\"emo_vec\":[0,0,0.55,0,0,0,0,0]}}",
	"sync_timeout": 30
}

Voice cloning — independent emotion reference audio

{
	"task": "/v1/ai_audio_engine/481991",
	"task_type": "formula",
	"init_images": [{
		"url": "http://action-public.meitudata.com/voice_clone/68c12c8a20160928587d19j69h5423.mp3",
		"profile": {
			"media_profiles": {
				"media_data_type": "url"
			},
			"version": "v1"
		}
	}],
	"params": "{\"rsp_media_type\":\"url\",\"parameter\":{\"task_type\":\"synth_once\",\"text\":\"今天天气真不错，你怎么知道的？\",\"emo_control_method\":2,\"emo_ref_path\":\"https://example.com/emo_ref.wav\",\"emo_weight\":0.8}}",
	"sync_timeout": 30
}

Response Fields

Generated results are periodically deleted. Download and save them promptly.

Field	Type	Description
request_id	string	Request identifier
trace_id	string	Trace ID
code	int	Business status code returned by the API. `0` indicates success; a nonzero value indicates failure.
error_code	int	Detailed error code (normally used with `code`); `0` indicates no error.
message	string	API message; normally `"success"` on success and an error description on failure.
tips	any	Additional tips; may be `null`
data	object	Business data

Fields in the data object

Field	Type	Description
status	int	Status: `-1` task not found; `0` created; `1` processing; `2` failed; `9` timed out—use the Task Query API to retrieve the result; `10` succeeded
result	object	Task result information
progress	number	Task progress
predict_elapsed	int	Estimated processing time in milliseconds
create_time	int64	Task creation timestamp in milliseconds
task_id	string	System-generated task ID that uniquely identifies the task
custom_task_id	string	User-defined task ID (if supplied when creating the task)
trace_id	string	Trace ID
client_info	string	Client information
init_images	object[]/null	Echoed input media

Fields in the result object

Field	Type	Description
id	string in the example; the source parameter table specifies int	Task ID
parameters	object	Task result information
data	object in the example; the source parameter table specifies int	The source parameter table describes predicted processing time in seconds; the response example uses a task metadata object.
msg	string in the example; the source parameter table specifies long	Task information
msg_id	string	Message ID

Fields in the parameters object

Field	Type	Description
code	int	`0` indicates a normal response; nonzero values indicate errors
name	String	The `user` parameter from the request
id	String	Job_ID
created_time	String	Task execution creation time
updated_time	String	Task execution completion time
msg	String	Task execution information
out_src	String	Generated audio file
text	String	Text to synthesize

Response Examples

Successful Response Example

Response status: 200

Content-Type: application/json; charset=utf-8

{
    "request_id": "",
    "trace_id": "",
    "code": 0,
    "error_code": 0,
    "message": "success",
    "tips": null,
    "data": {
        "status": 10,
        "result": {
            "id": "t_mt1a3i5n7bca14033a-6268-46ee-823e-c968beeb5b23",
            "parameters": {
                "clone_ori_path": "https://deepnet-baikal.obs.cn-north-4.myhuaweicloud.com/temporary/spliter/vocals_5c0e5136-ba21-42e9-b737-820e414bb0f1.wav",
                "code": 0,
                "create_time": "2025-09-30 17:34:09",
                "duration": 4.07,
                "id": "9939fbd5-7c74-47e1-9a71-d2130719bb71",
                "latency": 1.85,
                "msg": "success",
                "name": "0",
                "out_src": "https://deepnet-baikal.obs.cn-north-4.myhuaweicloud.com/temporary/zeroshot/9939fbd5-7c74-47e1-9a71-d2130719bb71_0.wav",
                "src": "https://action-public-release.obs.cn-north-4.myhuaweicloud.com/voice_clone/68c12c8a20160928587d19j69h5423.mp3?AccessKeyId=7YW1EYVPMUCUAFBLPDIN\u0026Expires=1759311244\u0026Signature=MF5vq%2B9y357Sjn23NmjWS5SHkf4%3D",
                "text": "You’re drinking energy drinks? Babe, it’s time to level up.",
                "ts_info": {

                },
                "updated_time": "2025-09-30 17:34:11"
            },
            "data": {
                "duration": {
                    "alg_process_time": 2087,
                    "created_timestamp": 1759224849,
                    "pull_timestamp": 1759224849,
                    "repost_time": 0,
                    "upload_time": 0,
                    "waiting_time": 78
                },
                "error_code": 0,
                "error_msg": "success",
                "extra": {

                },
                "media_info_list": [

                ],
                "msg_id": "a27d3c85-adb1-44a5-79ff-r0072bef0160",
                "parameter": {
                    "clone_ori_path": "https://deepnet-baikal.obs.cn-north-4.myhuaweicloud.com/temporary/spliter/vocals_5c0e5136-ba21-42e9-b737-820e414bb0f1.wav",
                    "code": 0,
                    "create_time": "2025-09-30 17:34:09",
                    "duration": 4.07,
                    "id": "9939fbd5-7c74-47e1-9a71-d2130719bb71",
                    "latency": 1.85,
                    "msg": "success",
                    "name": "0",
                    "out_src": "https://deepnet-baikal.obs.cn-north-4.myhuaweicloud.com/temporary/zeroshot/9939fbd5-7c74-47e1-9a71-d2130719bb71_0.wav",
                    "src": "https://action-public-release.obs.cn-north-4.myhuaweicloud.com/voice_clone/68c12c8a20160928587d19j69h5423.mp3?AccessKeyId=7YW1EYVPMUCUAFBLPDIN\u0026Expires=1759311244\u0026Signature=MF5vq%2B9y357Sjn23NmjWS5SHkf4%3D",
                    "text": "You’re drinking energy drinks? Babe, it’s time to level up.",
                    "ts_info": {

                    },
                    "updated_time": "2025-09-30 17:34:11"
                }
            },
            "msg": "success",
            "msg_id": "a27d3c85-adb1-44a5-79ff-r0072bef0160"
        },
        "progress": 1,
        "predict_elapsed": 10000,
        "create_time": 1759224844659,
        "task_id": "t_mt1a3i5n7bca14033a-6268-46ee-823e-c968beeb5b23",
        "custom_task_id": "",
        "trace_id": "e69535d2-23cd-40e3-bb27-addf15cd1457",
        "client_info": "",
        "init_images": null
    }
}

Pending Response Example

Use the Task Query API to retrieve the result. Response status: 200

Content-Type: application/json; charset=utf-8

{
    "request_id": "",
    "trace_id": "",
    "code": 0,
    "error_code": 0,
    "message": "success",
    "tips": null,
    "data": {
        "status": 9,
        "result": {
            "id": "t_mt1a3i5n7b3da6d589-46b5-4f66-a0bb-8dd22f2a172e"
        },
        "progress": 0,
        "predict_elapsed": 10000,
        "create_time": 1759202368761,
        "task_id": "t_mt1a3i5n7b3da6d589-46b5-4f66-a0bb-8dd22f2a172e",
        "custom_task_id": "",
        "trace_id": "9129a3a2-99c4-46ce-8731-0e100e2fbee7",
        "client_info": "",
        "init_images": null
    }
}

Failed Response Example

Response status: 400

Content-Type: application/json; charset=utf-8

{
    "request_id": "",
    "trace_id": "",
    "code": 20001,
    "error_code": 20001,
    "message": "ALGO_MODEL_CRASH",
    "tips": null,
    "data": {
        "status": 2,
        "result": {
            "id": "t_mt1a3i5n7be8d575cc-2ffb-4e0c-85a8-1824110e31b8",
            "code": 20001,
            "data": {
                "duration": {
                    "alg_process_time": 0,
                    "created_timestamp": 1759201989,
                    "pull_timestamp": 1759201989,
                    "repost_time": 0,
                    "upload_time": 0,
                    "waiting_time": 0
                },
                "error_code": 20001,
                "error_msg": "ALGO_MODEL_CRASH",
                "extra": {

                },
                "media_info_list": [

                ],
                "msg_id": "c1b09cb2-6e05-4d21-55ab-r007b1f21bec",
                "parameter": null
            },
            "msg": "ALGO_MODEL_CRASH",
            "msg_id": "c1b09cb2-6e05-4d21-55ab-r007b1f21bec"
        },
        "progress": 1,
        "predict_elapsed": 10000,
        "create_time": 1759201989444,
        "task_id": "t_mt1a3i5n7be8d575cc-2ffb-4e0c-85a8-1824110e31b8",
        "custom_task_id": "",
        "trace_id": "",
        "client_info": "",
        "init_images": null
    }
}

API-Specific Error Codes and Messages

ErrorCode	Error message	Description

General Error Codes

See API Error Codes

SDK Examples

Python

import json

import requests
from sign_sdk import sign

def api_call_example():
    key = "your_api_key"
    secret = "your_api_secret"
    url = "https://openapi.meitu.com/api/v1/sdk/sync/push"
    method = "POST"
    headers = {
        "Content-Type": "application/json",
        sign.HeaderHost: "openapi.meitu.com",
    }
    inner_params = {
        "rsp_media_type": "url",
        "parameter": {
            "task_type": "synth_once",
            "text": "我是你克隆的专属声音，快来用我制作视频吧。",
        },
    }
    payload = {
        "task": "/v1/ai_audio_engine/481991",
        "task_type": "formula",
        "init_images": [{
            "url": "http://action-public.meitudata.com/voice_clone/68c12c8a20160928587d19j69h5423.mp3",
            "profile": {
                "media_profiles": {"media_data_type": "url"},
                "version": "v1",
            },
        }],
        "params": json.dumps(inner_params, ensure_ascii=False),
        "sync_timeout": 30,
    }
    body = json.dumps(payload, ensure_ascii=False)
    signer = sign.Signer(key, secret)
    signed_request = signer.sign(url, method, headers, body)
    response = requests.Session().send(signed_request)
    print(f"Status: {response.status_code}")
    print(f"Response: {response.text}")

if __name__ == "__main__":
    api_call_example()

Go

package main

import (
	"fmt"
	"io"
	"net/http"

	"github.com/mtlab/api/signer"
)

func main() {
	key := "your_api_key"
	secret := "your_api_secret"
	signObj := signer.NewSigner(key, secret)
	url := "https://openapi.meitu.com/api/v1/sdk/sync/push"
	method := http.MethodPost
	headers := make(http.Header)
	headers.Set(signer.HeaderHost, "openapi.meitu.com")
	headers.Set("Content-Type", "application/json")
	body := `{
  "task": "/v1/ai_audio_engine/481991",
  "task_type": "formula",
  "init_images": [{
    "url": "http://action-public.meitudata.com/voice_clone/68c12c8a20160928587d19j69h5423.mp3",
    "profile": {
      "media_profiles": {"media_data_type": "url"},
      "version": "v1"
    }
  }],
  "params": "{\"rsp_media_type\":\"url\",\"parameter\":{\"task_type\":\"synth_once\",\"text\":\"我是你克隆的专属声音，快来用我制作视频吧。\"}}",
  "sync_timeout": 30
}`
	req, err := signObj.Sign(url, method, headers, body)
	if err != nil {
		fmt.Println("Failed to sign request:", err)
		return
	}
	resp, err := http.DefaultClient.Do(req)
	if err != nil {
		fmt.Println("Failed to send request:", err)
		return
	}
	defer resp.Body.Close()
	responseBody, err := io.ReadAll(resp.Body)
	if err != nil {
		fmt.Println("Read response failed:", err)
		return
	}
	fmt.Println("Response:", resp.StatusCode, string(responseBody))
}

PHP

<?php
require 'signer.php';

$key = 'your_api_key';
$secret = 'your_api_secret';
$signer = new Signer($key, $secret);
$url = 'https://openapi.meitu.com/api/v1/sdk/sync/push';
$method = 'POST';
$headers = ['Content-Type' => 'application/json'];
$innerParams = json_encode([
    'rsp_media_type' => 'url',
    'parameter' => [
        'task_type' => 'synth_once',
        'text' => '我是你克隆的专属声音，快来用我制作视频吧。',
    ],
]);
$body = json_encode([
    'task' => '/v1/ai_audio_engine/481991',
    'task_type' => 'formula',
    'init_images' => [[
        'url' => 'http://action-public.meitudata.com/voice_clone/68c12c8a20160928587d19j69h5423.mp3',
        'profile' => [
            'media_profiles' => ['media_data_type' => 'url'],
            'version' => 'v1',
        ],
    ]],
    'params' => $innerParams,
    'sync_timeout' => 30,
]);
$curl = $signer->sign($url, $method, $headers, $body);
$response = curl_exec($curl);
$status = curl_getinfo($curl, CURLINFO_HTTP_CODE);
if ($status === 0) {
    echo 'Error: ' . curl_error($curl);
} else {
    echo "Status: {$status}\n";
    echo "Response: {$response}\n";
}
curl_close($curl);
?>

Java

package com.meitu.openai.common;

import java.io.BufferedReader;
import java.io.InputStream;
import java.io.InputStreamReader;
import java.net.HttpURLConnection;
import java.net.URL;
import java.nio.charset.StandardCharsets;
import java.util.HashMap;
import java.util.Map;

public class Main {
    public static void main(String[] args) throws Exception {
        Signer signer = new Signer("your_api_key", "your_api_secret");
        String url = "https://openapi.meitu.com/api/v1/sdk/sync/push";
        String method = "POST";
        Map<String, String> headers = new HashMap<>();
        headers.put("Content-Type", "application/json");
        headers.put(Signer.HeaderHost, "openapi.meitu.com");
        String body = "{\n" +
                "  \"task\": \"/v1/ai_audio_engine/481991\",\n" +
                "  \"task_type\": \"formula\",\n" +
                "  \"init_images\": [{\n" +
                "    \"url\": \"http://action-public.meitudata.com/voice_clone/68c12c8a20160928587d19j69h5423.mp3\",\n" +
                "    \"profile\": {\n" +
                "      \"media_profiles\": {\"media_data_type\": \"url\"},\n" +
                "      \"version\": \"v1\"\n" +
                "    }\n" +
                "  }],\n" +
                "  \"params\": \"{\\\"rsp_media_type\\\":\\\"url\\\",\\\"parameter\\\":{\\\"task_type\\\":\\\"synth_once\\\",\\\"text\\\":\\\"我是你克隆的专属声音，快来用我制作视频吧。\\\"}}\",\n" +
                "  \"sync_timeout\": 30\n" +
                "}";
        Map<String, String> signedHeaders = signer.sign(url, method, headers, body);
        HttpURLConnection connection = (HttpURLConnection) new URL(url).openConnection();
        connection.setRequestMethod(method);
        for (Map.Entry<String, String> entry : signedHeaders.entrySet()) {
            connection.setRequestProperty(entry.getKey(), entry.getValue());
        }
        connection.setDoOutput(true);
        connection.getOutputStream().write(body.getBytes(StandardCharsets.UTF_8));
        int status = connection.getResponseCode();
        InputStream stream = status >= 400
                ? connection.getErrorStream()
                : connection.getInputStream();
        try (BufferedReader reader = new BufferedReader(new InputStreamReader(stream))) {
            StringBuilder response = new StringBuilder();
            String line;
            while ((line = reader.readLine()) != null) {
                response.append(line);
            }
            System.out.println("Response: " + status + " " + response);
        }
    }
}

General-Purpose Speech Multitask Engine

Description

Version

Media Requirements

API URL

Method

Authentication

Request Parameters

1. lufsr (section name in the source document; the task_type value is lufs, which retrieves audio loudness)

2. trans_audio (speech transcription and translation)

3. ducking (vocal-ducking keyframes)

4. train (voice training and voice-profile storage)

5. synth (generation using an existing voice profile)

6. synth_once (one-time cloning and generation without saving the voice profile)

Emotion Control

7. edit (voice editing)

8. trans (translated generation using an existing voice profile)

9. trans_once (real-time voice extraction and translation)

10. conversion (voice conversion)

Request Examples

Response Fields

Response Examples

Successful Response Example

Pending Response Example

Failed Response Example

API-Specific Error Codes and Messages

General Error Codes

SDK Examples

Python

Go

PHP

Java

1. lufsr (section name in the source document; the `task_type` value is `lufs`, which retrieves audio loudness)