Create Async Speech Synthesis Task | Unisound Maas Platform

Guide API Reference

Skill Document Pricing Token Plan Releases

Create Async TTS Task

Create an async long-text TTS task, supporting direct text input or text file for synthesis.

POST/v1/audio/speech/tasks

Authorization

AuthorizationstringheaderRequired

HTTP: Bearer Auth

Security Scheme Type: http
HTTP Authorization Scheme: Bearer API_key，Used to verify account information, can be viewed in Metered & Packs > API Key Management .

Request Header

Content-Typeenum<string>Default:application/jsonRequired

The media type of the request body, please set it toapplication/jsonto ensure the correct format of the requested data.

Available options:application/json

Request Bodyapplication/json

modelstringRequired

Model code, optional values: u2-tts, u2-tts-clone, u2-tts-design

textstring

The text to be synthesized. When model is u2-tts, the maximum is 50,000 characters; when model is u2-tts-clone or u2-tts-design, the maximum is 20,000 characters. Either this or text_file_id is required.

text_file_idlong

The text file ID (txt) to be synthesized. When model is u2-tts, max 50,000 chars; when model is u2-tts-clone or u2-tts-design, max 20,000 chars. Either this or text is required.

voice_settingobjectRequired

Basic voice settings

voice_setting.voice_idstring

System/clone/design voice ID, can be obtained through the query available voices API

voice_setting.speedinteger

Speed range [0, 100], default 50.

See Voice Parameter Support for details.

voice_setting.volumeinteger

Volume range [0, 100], default 50.

See Voice Parameter Support for details.

voice_setting.pitchinteger

Pitch range [0, 100], default 50.

See Voice Parameter Support for details.

voice_setting.brightinteger

Brightness range [50, 100], default 50.

See Voice Parameter Support for details.

voice_setting.emotionstring

Pronunciation emotion. Optional values: happy, angry, depressed, whisper, loudly, neutral, corresponding to 6 emotions: happy, angry, depressed, whisper, loudly, and neutral. Currently only supported by the cn_male_chenyu speaker when language is Chinese and dialect is default.

voice_setting.languagestring

Pronunciation language. Optional values: zh, ja, ko, th, vi, id, ms, my, lo, km, tl, en.

Supported languages by voice:

`cn_female_shasha`: `zh` (Chinese), `ja` (Japanese), `ko` (Korean), `th` (Thai), `vi` (Vietnamese), `id` (Indonesian), `km` (Khmer)
`cn_female_jiajia`: `zh` (Chinese), `id` (Indonesian), `ms` (Malay), `my` (Burmese), `lo` (Lao)
`cn_male_chenyu`: `zh` (Chinese), `tl` (Filipino)
`en_male_johnny`, `en_female_jane`: `en` (English)
Other system voices: `zh` (Chinese)

voice_setting.dialectstring

Pronunciation dialect. Optional values: `yueyu`, `sichuan`.

Currently, only the `cn_male_chenyu` voice supports the following two dialects:

`yueyu` (Cantonese)
`sichuan` (Sichuan dialect)

audio_settingobject

Audio output settings

audio_setting.audio_sample_rateinteger

Sample rate, enum [8000, 16000, 24000, 32000], default 32000

audio_setting.formatstring

Output format, enum [mp3, pcm], default mp3

audio_setting.channelinteger

Number of channels, enum [1]

pronunciation_dictobject

Custom pronunciation rules

pronunciation_dict.tonestring[ ]

Pronunciation/phonetic replacement rules, example: ["Liangshan/Liang<py>liang2</py>shan"].

See Custom Pronunciation Rules for tag details and support.

Response Body Structure

task_idstring

Unique identification ID of the current async synthesis task

file_idlong

Audio file ID returned when the task is successfully created

usage_charactersinteger

Number of characters consumed

base_respobject

base_resp.status_codeinteger

Request status code, 0 = normal

base_resp.status_msgstring

Status description, success indicates success

Create T2A Async Task

curl --request POST \
  --url https://maas-api.unisound.com/v1/audio/speech/tasks \
  --header 'Authorization: Bearer <api_key>' \
  --header 'Content-Type: application/json;charset=UTF-8' \
  --data '{
  "model": "u2-tts",
  "text": "The story of the heroes of Liangshan is widely known",
  "voice_setting": {
    "voice_id": "cn_male_chenyu",
    "volume": 50,
    "pitch": 50,
    "bright": 50,
    "emotion": "happy",
    "language": "zh"
  },
  "pronunciation_dict": {
    "tone": [
      "Liangshan/Liang<py>liang2</py>shan"
    ]
  },
  "audio_setting": {
    "audio_sample_rate": 16000,
    "format": "mp3"
  }
}'

200

{
  "task_id": "95157322514444",
  "file_id": 871009177767936,
  "usage_characters": 11,
  "base_resp": {
    "status_code": 0,
    "status_msg": "success"
  }
}

Voice Parameter Support

Support for speed, volume, pitch, and brightness by voice ID, language, and dialect:

Voice ID	Language	Dialect	speed	volume	pitch	bright
cn_female_xiaodi_warm	zh	default	✓	✓	✓	✓
cn_male_chenyu	zh	default	✗	✓	✓	✓
cn_male_chenyu	zh	yueyu	✓	✓	✓	✓
cn_male_chenyu	zh	sichuan	✓	✓	✓	✓
cn_male_chenyu	tl	default	✗	✗	✗	✗
en_male_johnny	en	default	✓	✓	✓	✓
en_female_jane	en	default	✓	✓	✓	✓
cn_male_chenyu_robot	zh	default	✗	✓	✓	✓
cn_male_chenyu_elderly	zh	default	✗	✓	✓	✓
cn_male_chenyu_steady	zh	default	✗	✓	✓	✓
cn_male_chenyu_fluent	zh	default	✗	✓	✓	✓
cn_female_shasha	zh, ja, ko, th, vi, id	default	✗	✓	✓	✓
cn_female_shasha	km	default	✗	✗	✗	✗
cn_female_jiajia	zh	default	✗	✓	✓	✓
cn_female_jiajia	id, ms, my, lo	default	✗	✗	✗	✗
cn_female_ruolin	zh	default	✗	✓	✓	✓
cn_male_jingming	zh	default	✗	✗	✗	✗
cn_female_shasha_intellectual	zh	default	✗	✗	✗	✗
cn_female_jiajia_intellectual	zh	default	✗	✗	✗	✗

Custom Pronunciation Rules

Use the following tags to control pronunciation and segmentation. Tag support varies by voice; see the support table below.

Category	Tag	Description	Example	Effect
Number	<value>	Pronounce the digit string as a numeric value	这是第<value>110</value>会议室	这是第一百一十会议室
	<code>	Pronounce the digit string as a code (“1” as “一”)	这是第<code>110</code>会议室	这是一一零会议室
	<tel>	Pronounce the digit string as a phone number (“1” as “幺”)	会议室分机号为<tel>110</tel>	会议室分机号为幺幺零
Phonetics	<py>	Force the engine to pronounce Chinese characters with the specified pinyin	搜索结果为<py>wei2</py>空<py>kong1</py>	搜索结果为(wei2)空(kong1)
Phonetics	<pname>	Correctly pronounce polyphonic characters used as surnames	播放<pname>单田芳</pname>的评书	播放单(shan4)田芳的评书
Segmentation	<word>	Force word segmentation to fix incorrect splits, e.g. distinguish “(乒乓球)(拍卖)(完了)” from “(乒乓)(球拍)(卖完了)”	<word>乒乓</word><word>球拍</word><word>卖完了</word>	(乒乓)(球拍)(卖完了)
Segmentation	<phrase>	Force phrase breaks (longer pause than word) to fix incorrect phrasing, e.g. distinguish “(爸爸亲了我妈妈)(也亲了我)” from “(爸爸亲了我)(妈妈也亲了我)”	<phrase>爸爸亲了我</phrase><phrase>妈妈也亲了我</phrase>	(爸爸亲了我) (妈妈也亲了我)
English	<letter>	Spell the marked English text letter by letter	世界卫生组织的英文缩写<letter>WHO</letter>	世界卫生组织的英文缩写是(W)(H)(O)
Other	<mute>	Insert a pause of the specified duration (milliseconds)	请您再说一遍<mute>300</mute>或说取消	请您再说一遍(pause 300ms)或说取消
Other	<sub>	Replace the marked text with the alias attribute value when speaking	气压的测量单位是<sub alias="毫米汞柱">mmHg</sub>	气压的测量单位是毫米汞柱

Tag Support

Voice ID	Language	Dialect	Supported Tags
cn_female_xiaodi_warm	zh	default	All supported
cn_male_chenyu	zh	default	Partial (value、code、tel、py、letter、sub、word)
cn_male_chenyu	zh	yueyu	All supported
cn_male_chenyu	zh	sichuan	All supported
cn_male_chenyu	tl	default	Not supported
en_male_johnny	en	default	All supported
en_female_jane	en	default	All supported
cn_male_chenyu_robot	zh	default	Partial (value、code、tel、py、letter、sub、word)
cn_male_chenyu_elderly	zh	default	Partial (value、code、tel、py、letter、sub、word)
cn_male_chenyu_steady	zh	default	Partial (value、code、tel、py、letter、sub、word)
cn_male_chenyu_fluent	zh	default	Partial (value、code、tel、py、letter、sub、word)
cn_female_shasha	zh, ja, ko, th, vi, id	default	Partial (value、code、tel、py、letter、sub、word)
cn_female_shasha	km	default	Not supported
cn_female_jiajia	zh	default	Partial (value、code、tel、py、letter、sub、word)
cn_female_jiajia	id, ms, my, lo	default	Not supported
cn_female_ruolin	zh	default	Partial (value、code、tel、py、letter、sub、word)
cn_male_jingming	zh	default	Not supported
cn_female_shasha_intellectual	zh	default	Not supported
cn_female_jiajia_intellectual	zh	default	Not supported