SuperAPI AI Directory

doubao-tts

General & other

View repository

Updated 11-13 22:14

About this skill

Doubao Speech Synthesis Skill (Doubao TTS)

This skill uses the Volcano Engine Doubao speech synthesis large model'sone-way streaming HTTP V3 interfaceto synthesize text into speech audio files.

Official documentation:Doubao Speech Synthesis API documentation

Prerequisites

The user needs to provide the following environment variables (set via export or passed in as script parameters):

  • DOUBAO_APP_ID: the APP ID obtained from the Volcano Engine console
  • DOUBAO_ACCESS_KEY: the Access Token obtained from the Volcano Engine console
  • DOUBAO_SPEAKER: timbre ID (optional, defaults to zh_female_xiaohe_uranus_bigtts)
    • Voice clone timbres start with S_ and can be created in the official backend:Timbre Clone Console
    • See the timbre comparison table below for official preset timbres

If the user has not set DOUBAO_APP_ID and DOUBAO_ACCESS_KEY,first remind the user to set them, and tell them how to obtain them: log in to the Volcano Engine console → Doubao Voice → create an application → obtain the APP ID and Access Token.

If the user has not set DOUBAO_SPEAKER, use zh_female_xiaohe_uranus_bigttsby default (Xiaohe 2.0, general Chinese female voice).

Usage flow

1. Confirm parameters

Confirm the following information with the user (items with reasonable defaults can skip confirmation):

Parameter Description Default value
Text to synthesize The text content to convert to speech (required)
Timbre ID (speaker) Timbre identifier. Voice clone timbres start with S_ ; select an appropriate timbre from the timbre table below according to the context $DOUBAO_SPEAKER or zh_female_xiaohe_uranus_bigtts
Resource ID (resource_id) For voice clone use seed-icl-1.0; for official 1.0 timbres use seed-tts-1.0; for official 2.0 timbres use seed-tts-2.0 seed-tts-2.0
Audio format (format) mp3 / ogg_opus / pcm mp3
Sample rate (sample_rate) Options: 8000/16000/22050/24000/32000/44100/48000 24000
Emotion (emotion) Emotion type, supported only by multi-emotion timbres. Chinese timbres:happy / sad / angry / surprised / fear / hate / excited / coldness / neutral , etc.; English timbres:neutral / happy / angry / sad / excited / chat / warm , etc. Not set
Emotion intensity (emotion_scale) Range 1~5; the larger the value, the more obvious the emotion (non-linear); must be used together with emotion Use 4
Output file name File name of the generated audio output.mp3

2. Execute synthesis

Run the script:

# 基础用法(resource-id 自动推断)
python3 /path/to/doubao-tts/scripts/tts_synthesize.py \
  --text "要合成的文本" \
  --speaker "${DOUBAO_SPEAKER:-zh_female_xiaohe_uranus_bigtts}" \
  --format mp3 \
  --sample-rate 24000 \
  --output /mnt/user-data/outputs/output.mp3

# 带情感的合成(仅多情感音色支持)
python3 /path/to/doubao-tts/scripts/tts_synthesize.py \
  --text "要合成的文本" \
  --speaker "zh_female_shuangkuaisisi_emo_v2_mars_bigtts" \
  --emotion happy \
  --emotion-scale 4 \
  --output /mnt/user-data/outputs/output.mp3

Environment variables DOUBAO_APP_ID and DOUBAO_ACCESS_KEY must already be set. They can also be passed directly via the --app-id and --access-key parameters.

3. Output result

The script saves the synthesized audio to the specified path. After synthesis is complete, use the present_files tool to present the file to the user.

Voice Comparison Table

The Agent should select the most appropriate voice ID from the table below based on the context (the role, scenario, and style requested by the user).

For the complete voice list, see the official documentation:Voice List

Doubao Speech Synthesis 2.0 voices (recommended, resource_id uses seed-tts-2.0)

Scenario Voice Name voice_type (Voice ID) Language Applicable Role/Description
General Xiaohe 2.0 ⭐Default zh_female_xiaohe_uranus_bigtts Chinese General female voice, natural and friendly
General Vivi 2.0 zh_female_vv_uranus_bigtts Chinese/Japanese/Indonesian/Spanish Multilingual female voice, emotionally expressive
General Yunzhou 2.0 zh_male_m191_uranus_bigtts Chinese General male voice
General Xiaotian 2.0 zh_male_taocheng_uranus_bigtts Chinese General male voice, young
General Liu Fei 2.0 zh_male_liufei_uranus_bigtts Chinese General male voice
General Charming Sophie 2.0 zh_male_sophie_uranus_bigtts Chinese General male voice
General Fresh Female Voice 2.0 zh_female_qingxinnvsheng_uranus_bigtts Chinese Fresh and elegant female voice
General Sweet Xiaoyuan 2.0 zh_female_tianmeixiaoyuan_uranus_bigtts Chinese Sweet female voice
General Sweet Taozi 2.0 zh_female_tianmeitaozi_uranus_bigtts Chinese Sweet female voice
General Cheerful Sisi 2.0 zh_female_shuangkuaisisi_uranus_bigtts Chinese Cheerful and lively female voice
General Girl Next Door 2.0 zh_female_linjianvhai_uranus_bigtts Chinese Friendly girl-next-door style female voice
General Youthful Zixin 2.0 zh_male_shaonianzixin_uranus_bigtts Chinese Youthful male voice
General Charming Girlfriend 2.0 zh_female_meilinvyou_uranus_bigtts Chinese Charming mature female voice
General Fluent Female Voice 2.0 zh_female_liuchangnv_uranus_bigtts Chinese Fluent and natural female voice, suitable for reading long texts aloud
General Refined Yichen 2.0 zh_male_ruyayichen_uranus_bigtts Chinese Refined and mature male voice
Role-play Intellectual Cancan 2.0 zh_female_cancan_uranus_bigtts Chinese Intellectual and capable female voice
Role-play Coy Junior Schoolgirl 2.0 zh_female_sajiaoxuemei_uranus_bigtts Chinese Coy and cute female voice
Video Dubbing Monkey King 2.0 zh_male_sunwukong_uranus_bigtts Chinese Sun Wukong character, heroic
Video Dubbing Peppa Pig 2.0 zh_female_peiqi_uranus_bigtts Chinese Cartoon child character
Video Dubbing Dayi 2.0 zh_male_dayi_uranus_bigtts Chinese Video dubbing male voice
Video dubbing Black Cat Detective Agency Mizai 2.0 zh_female_mizai_uranus_bigtts Chinese Cute cartoon female voice
Video dubbing Chicken Soup Female 2.0 zh_female_jitangnv_uranus_bigtts Chinese Warm inspirational female voice
Education scenario Teacher Tina 2.0 zh_female_yingyujiaoxue_uranus_bigtts Chinese/British English Teacher/education scenario
Customer service scenario Warm Sun Female Voice 2.0 zh_female_kefunvsheng_uranus_bigtts Chinese Dedicated for customer service, warm and professional
Audiobook reading Children's Picture Book 2.0 zh_female_xiaoxue_uranus_bigtts Chinese Children's stories, picture book reading
Multilingual Tim en_male_tim_uranus_bigtts American English English male voice
Multilingual Dacey en_female_dacey_uranus_bigtts American English English female voice
Multilingual Stokie en_female_stokie_uranus_bigtts American English English female voice

Doubao Speech Synthesis 1.0 voices (resource_id uses seed-tts-1.0)

General scenario

Scenario Voice name voice_type (voice ID) Language Applicable role/description
General Cancan/Shiny zh_female_cancan_mars_bigtts Chinese/American English Lively and cheerful female voice
General Vivi zh_female_vv_mars_bigtts Chinese Standard female voice
General Straightforward Sisi/Skye zh_female_shuangkuaisisi_moon_bigtts Chinese/American English Bright and lively female voice
General Warm Ahu/Alvin zh_male_wennuanahu_moon_bigtts Chinese/American English Warm and friendly male voice
General Youth Zixin/Brayan zh_male_shaonianzixin_moon_bigtts Chinese/American English Youthful boy voice
General Fresh female voice zh_female_qingxinnvsheng_mars_bigtts Chinese Fresh and elegant
General Intellectual female voice zh_female_zhixingnvsheng_mars_bigtts Chinese Intellectual and mature female voice
General Refreshing male college student zh_male_qingshuangnanda_mars_bigtts Chinese Refreshing male college student voice
General Girl next door zh_female_linjianvhai_moon_bigtts Chinese Friendly girl-next-door style
General Erudite uncle zh_male_yuanboxiaoshu_moon_bigtts Chinese Knowledgeable uncle feel
General Sunny youth zh_male_yangguangqingnian_moon_bigtts Chinese Positive and sunny male voice
General Sweet Xiaoyuan zh_female_tianmeixiaoyuan_moon_bigtts Chinese Sweet female voice
General Sweet Yueyue zh_female_tianmeiyueyue_moon_bigtts Chinese Sweet female voice
General Clear Zizi zh_female_qingchezizi_moon_bigtts Chinese Clear and clean female voice
General Cheerful Sister zh_female_kailangjiejie_moon_bigtts Chinese Cheerful Big Sister
General Boy Next Door zh_male_linjiananhai_moon_bigtts Chinese Friendly Boy-Next-Door Male Voice
General Inspirational Chicken Soup zh_female_xinlingjitang_moon_bigtts Chinese Warm and Inspirational Female Voice
General Happy Xiaodong zh_male_xudong_conversation_wvae_bigtts Chinese Male Voice for Dialogue Scenarios
General Friendly Female Voice zh_female_qinqienvsheng_moon_bigtts Chinese Friendly and Gentle Female Voice
Education Teacher Tina zh_female_yingyujiaoyu_mars_bigtts Chinese/British English Teaching Scenario

Multi-emotion Voices (support emotion parameter)

Voice Name voice_type (Voice ID) Supported Emotions Applicable Scenarios
Frank Sisi (Multi-emotion) zh_female_shuangkuaisisi_emo_v2_mars_bigtts Happy/Sad/Angry/Surprised/Excited/Indifferent/Neutral Emotionally rich dialogue, stories
Gentle Girlfriend (Multi-emotion) zh_female_roumeinvyou_emo_v2_mars_bigtts Happy/Sad/Angry/Surprised/Fear/Disgust/Excited/Indifferent/Neutral Emotionally rich female voice
Aloof Queen (Multi-emotion) zh_female_gaolengyujie_emo_v2_mars_bigtts Happy/Sad/Angry/Surprised/Fear/Disgust/Excited/Indifferent/Neutral Queen/Aloof female character
Cold Older Brother (Multi-emotion) zh_male_lengkugege_emo_v2_mars_bigtts Angry/Indifferent/Fear/Happy/Disgust/Neutral/Sad/Depressed Cold/Aloof male character
Tsundere CEO (Multi-emotion) zh_male_aojiaobazong_emo_v2_mars_bigtts Neutral/Happy/Angry/Disgust Domineering CEO, assertive male
Refined Boyfriend (Multi-emotion) zh_male_ruyayichen_emo_v2_mars_bigtts Happy/Sad/Angry/Fear/Excited/Indifferent/Neutral Gentle and refined boyfriend
Handsome Boyfriend (Multi-emotion) zh_male_junlangnanyou_emo_v2_mars_bigtts Happy/Sad/Angry/Surprised/Fear/Neutral Handsome boyfriend
Sunny Youth (Multi-emotion) zh_male_yangguangqingnian_emo_v2_mars_bigtts Happy/Sad/Angry/Fear/Excited/Indifferent/Neutral Sunny and positive youth
Gentle Young Master (Multi-emotion) zh_male_yourougongzi_emo_v2_mars_bigtts Happy/Angry/Fear/Disgust/Excited/Neutral/Depressed Gentle and elegant male
Beijing Young Master (Multi-emotion) zh_male_beijingxiaoye_emo_v2_mars_bigtts Angry/Surprised/Fear/Excited/Indifferent/Neutral Beijing accent male voice
Guangzhou De Ge (Multi-emotion) zh_male_guangzhoudege_emo_mars_bigtts Angry/Fear/Neutral Guangzhou dialect male voice
Beijing Chatterbox (Multi-emotion) zh_male_jingqiangkanye_emo_mars_bigtts Happy/Angry/Surprised/Disgust/Neutral Beijing Chatterbox style
Sweetie Xiaomei (Multi-emotion) zh_female_tianxinxiaomei_emo_v2_mars_bigtts Sad/Fear/Disgust/Neutral Sweet female voice
Neighbor Auntie (Multi-emotion) zh_female_linjuayi_emo_v2_mars_bigtts Neutral/Angry/Indifferent/Depressed/Surprised Middle-aged woman
Late-Night Podcast (Multi-emotion) zh_male_shenyeboke_emo_v2_mars_bigtts Surprised/Sad/Neutral/Disgust/Happy/Fear/Excited/Depressed/Indifferent/Angry Late-night podcast/radio
Candice en_female_candice_emo_v2_mars_bigtts Deep affection/Angry/ASMR/Conversational/Excited/Cheerful/Neutral/Warm English female voice
Serena en_female_skye_emo_v2_mars_bigtts Deep affection/Angry/ASMR/Conversational/Excited/Cheerful/Neutral/Sad/Warm English female voice
Glen en_male_glen_emo_v2_mars_bigtts Deep affection/Angry/ASMR/Conversational/Excited/Cheerful/Neutral/Sad/Warm English male voice
Sylus en_male_sylus_emo_v2_mars_bigtts Deep affection/Angry/ASMR/Authoritative/Conversational/Excited/Cheerful/Neutral/Sad/Warm English male voice
Corey en_male_corey_emo_v2_mars_bigtts Angry/ASMR/Authoritative/Conversational/Deep affection/Excited/Cheerful/Neutral/Sad/Warm British English male voice
Nadia en_female_nadia_tips_emo_v2_mars_bigtts Deep affection/Angry/ASMR/Conversational/Excited/Cheerful/Neutral/Sad/Warm British English female voice

Fun accents

Voice name voice_type (voice ID) Accent/Language Description
Taiwan Xiao He zh_female_wanwanxiaohe_moon_bigtts Taiwan Mandarin Taiwan-accented female voice
Beijing Chatty Guy/Harmony zh_male_jingqiangkanye_moon_bigtts Beijing accent Beijing-accented chatty guy
Beijing Young Master zh_male_beijingxiaoye_moon_bigtts Beijing accent Beijing-accented male voice
Adorable Sichuan Girl zh_female_daimengchuanmei_moon_bigtts Sichuan accent Sichuan-flavored female voice
Yuzhou Zixuan zh_male_yuzhouzixuan_moon_bigtts Henan accent Henan-accented male voice
Guangxi Yuanzhou zh_male_guangxiyuanzhou_moon_bigtts Guangxi accent Guangxi-accented male voice
Guangzhou De Ge zh_male_guozhoudege_moon_bigtts Guangdong accent Guangdong-accented male voice
Bay Area Uncle zh_female_wanqudashu_moon_bigtts Guangdong accent Guangdong accent
Haoyu Xiao Ge zh_male_haoyuxiaoge_moon_bigtts Qingdao accent Qingdao-accented male voice
Meituo Jie'er zh_female_meituojieer_moon_bigtts Changsha accent Changsha-accented female voice
Cantonese Xiao Tang zh_female_yueyunv_mars_bigtts Cantonese Cantonese female voice

Role-play

Voice name voice_type (voice ID) Role characteristics
Cute Milky Kid zh_male_naiqimengwa_mars_bigtts Cute kid, child role
Grandma zh_female_popo_mars_bigtts Elderly female, grandmother
Aloof Queen zh_female_gaolengyujie_moon_bigtts Aloof queen style
Tsundere CEO zh_male_aojiaobazong_moon_bigtts Domineering CEO, tsundere male
Charming Girlfriend zh_female_meilinvyou_moon_bigtts Charming girlfriend
Gentle Girlfriend zh_female_sajiaonvyou_moon_bigtts Gentle, clingy girlfriend
Clingy Junior Girl zh_female_yuanqinvyou_moon_bigtts Clingy, cute junior girl
Dongfang Haoran zh_male_dongfanghaoran_moon_bigtts Righteous male lead
Late-Night Podcast zh_male_shenyeboke_moon_bigtts Late-night radio host

IP voice imitation

Voice name voice_type (voice ID) Character/Prototype
Zhu Bajie zh_male_zhubajie_mars_bigtts Zhu Bajie from Journey to the West
Tang Seng zh_male_tangseng_mars_bigtts Tang Seng from Journey to the West
Luban No. 7 zh_male_lubanqihao_mars_bigtts Luban No. 7 from Honor of Kings
Kasukabe Sister zh_female_jiyejizi2_mars_bigtts Doraemon character
Female Thor zh_female_leidian_mars_bigtts Thor character
Zhuang Zhou zh_male_zhuangzhou_mars_bigtts Honor of Kings Zhuang Zhou

Agent Voice Selection Guide

Intelligently select a voice based on context:

Use Case Recommended Voice Voice ID
Default/General Xiaohe 2.0 zh_female_xiaohe_uranus_bigtts
Formal/News Broadcasting Intellectual Female Voice zh_female_zhixingnvsheng_mars_bigtts
Children's Stories/Picture Books Children's Picture Book 2.0 zh_female_xiaoxue_uranus_bigtts
English Content Dacey en_female_dacey_uranus_bigtts
Late Night/Radio/Podcast Late Night Podcast (Multi-Emotion) zh_male_shenyeboke_emo_v2_mars_bigtts
Audiobooks/Long-form Reading Fluent Female Voice 2.0 zh_female_liuchangnv_uranus_bigtts
Customer Service/Assistant Warm Sun Female Voice 2.0 zh_female_kefunvsheng_uranus_bigtts
Education/Teaching Teacher Tina 2.0 zh_female_yingyujiaoxue_uranus_bigtts
Taiwanese Accent Wanwan Xiaohe zh_female_wanwanxiaohe_moon_bigtts
Beijing Accent Jingqiang Kanye zh_male_jingqiangkanye_moon_bigtts
Ancient Style/Wuxia Characters Dongfang Haoran zh_male_dongfanghaoran_moon_bigtts
Domineering CEO Tsundere CEO zh_male_aojiaobazong_moon_bigtts
Cute Child Baby-voiced Child zh_male_naiqimengwa_mars_bigtts
Emotional Dialogue Cheerful Sisi (Multi-Emotion) zh_female_shuangkuaisisi_emo_v2_mars_bigtts
User-defined Custom Cloned Voice User-provided S_ voice S_xxxxxxxx(User voice clone ID)

Important Notes

  • Voice clone voices(starting with S_) must use seed-icl-1.0 as the resource_id, and cannot use seed-tts-1.0 or seed-tts-2.0
  • 2.0 voices(uranus series) must use seed-tts-2.0, and do not support the legacy V1 interface
  • 1.0 voices(mars/moon series) use seed-tts-1.0
  • If the text provided by the user is very long (over 5000 characters), it is recommended to synthesize it in segments and then concatenate
  • The script uses requests.Session to implement connection reuse, in line with official best practices
  • The audio data returned by the streaming interface is base64-encoded, and the script will automatically decode and concatenate it
  • If the user requests adjustments to speech rate, volume, pitch, etc., the corresponding parameters can be passed in

Error Handling

  • 40402003: The text is too long and needs to be segmented
  • 40000001: Parameter error; check whether the voice ID and resource_id match
  • 40300001: Authentication failed; check the APP ID and Access Key
  • If a quota-related error occurs: the trial version usage has been exhausted, and the official version needs to be activated