Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

Get Started Free →

Doubao Open TTS · LeoYeAI | DecimalAI

Doubao Open Tts

SkillScore70Maintained

doubao-open-tts

Text-to-Speech service using Doubao (Volcano Engine) API with 200+ voices, interactive voice selection, and multilingual support

design-mediaby LeoYeAI

Links this skill to your agents and keeps it current with the author’s updates.

Automatic

AI Engineer pack

1 · One-time — install the CLI (docs ↗)

pip install decimalai

2 · Get the skill — writes .claude/skills/doubao-open-tts/SKILL.md

decimalai skills pull doubao-open-tts --out .claude/skills/

No CLI? Get the same file — curl · raw file

curl -fsSL --create-dirs -o .claude/skills/doubao-open-tts/SKILL.md https://app.decimal.ai/s/doubao-open-tts/SKILL.md

Raw file: /s/doubao-open-tts/SKILL.md ↗ · machine-readable summary at /s/doubao-open-tts.json ↗

In productionGlobal

how this skill behaves on real traffic · last 30 days

No production data for this skill yet — usage numbers appear here once agents start running it.

Anonymized aggregate across every org using this skill — proof from real production use, not a demo.

Does this skill help? · Benchmark: with vs. withoutsign in to run this on your agents

23 test cases from the publisher’s eval.yaml · each run twice on gemini-3.6-flash — with the skill injected, then without · last run 56d ago

Yes — +57% lift · at higher cost+390% tokens0% turns70% of whole cases pass with skill (16/23)

Test case	Without → With	Effect	Δ tokens	Δ turns
case-19	✗→✓	▲ Improved	—	—
case-11	✗→✓	▲ Improved	—	—
case-21	✗→✓	▲ Improved	—	—
case-09	✗→✓	▲ Improved	—	—
case-07	✗→✓	▲ Improved	—	—

Showing 5 of 23 cases

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v4 — 7/30/2026

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

LeoYeAI / openclaw-master-skills

skills/doubao-api-open-tts/SKILL.md

View on GitHub ↗

License MITImported 8/2/2026Version v1Bundled 2 files · scripts

Skill contents

scripts/

SKILL.md

Doubao Open TTS

Text-to-Speech (TTS) service using Doubao (Volcano Engine) API V1 interface to convert text into natural-sounding speech.

Features

🎙️ 200+ Voice Options - Default: Shiny (灿灿) for general scenarios
🔊 Multiple Audio Formats - Supports mp3, pcm, wav
⚡ Adjustable Parameters - Speed and volume control
📦 Dual Interface - Command-line tool + Python API
🎯 Voice Categorization - Browse voices by category
🔐 Mandatory API Configuration - Requires API credentials; stops and prompts user if not configured (default voice only available with explicit user confirmation)

Quick Start for Agents

重要提示: 当检测到没有API配置时，Agent必须立即停止并询问用户是否配置API，不允许直接使用默认音色生成音频，除非用户明确坚持使用默认音色。

When using this SKILL, the Agent should follow this workflow:

Step 1: Check API Configuration

First, check if API credentials are configured:

python
from scripts.tts import check_api_config, setup_api_config

# Check if API is configured
config = check_api_config()
if not config:
    # API not configured, MUST stop and ask user
    # DO NOT proceed with default voice
    print("⚠️ API credentials not found. Cannot proceed without API configuration.")
    # ... ask user for credentials or if they want to use default

Step 2: Handle Missing API Configuration

如果API凭证缺失，Agent必须停止并询问用户: 用户可以选择配置API，或者明确坚持使用默认音色（仅当用户明确要求时才允许）。

Agent: I need API credentials to use the Doubao TTS service. 
       Please provide the following information:

1. App ID (APPID)
2. Access Token
3. Secret Key

You can get these from: https://console.volcengine.com/

Please enter your App ID:

User: 4269953601

Agent: Please enter your Access Token:

User: uAGE3iP8nJf3ewu-d6U1P6Jthv7i1DH7

Agent: Please enter your Secret Key:

User: Dfta36aYgcdqY-ylHoh9FQyqiERS5iTv

Agent: [Saves credentials to .env file]
       ✅ API configuration saved successfully!

用户选择: 如果用户不想配置API但坚持使用默认音色:

User: 我不想配置API，就用默认音色吧

Agent: ⚠️ 确认使用默认音色吗？这将使用内置的默认音色生成音频。
       输入 'yes' 确认使用默认音色，或提供API凭证以获得更好的体验。

User: yes

Agent: [继续执行，使用默认音色]

Step 3: Use the Service

After API is configured OR user explicitly confirmed to use default voice:

python
from scripts.tts import VolcanoTTS

tts = VolcanoTTS()
output = tts.synthesize("Hello world", output_file="output.mp3")

API Configuration Detection

Function: `check_api_config()`

Checks if API credentials are available. Returns config dict or None.

python
from scripts.tts import check_api_config

config = check_api_config()
if config:
    print(f"App ID: {config['app_id']}")
    print(f"Access Token: {config['access_token'][:10]}...")
    print(f"Secret Key: {config['secret_key'][:10]}...")
else:
    print("API not configured")

Function: `setup_api_config(app_id, access_token, secret_key, voice_type=None)`

Saves API credentials to the .env file in the SKILL directory.

python
from scripts.tts import setup_api_config

# Save credentials
setup_api_config(
    app_id="4269953601",
    access_token="uAGE3iP8nJf3ewu-d6U1P6Jthv7i1DH7",
    secret_key="Dfta36aYgcdqY-ylHoh9FQyqiERS5iTv",
    voice_type="zh_female_cancan_mars_bigtts"  # optional
)

print("✅ Configuration saved to .env file")

Complete Agent Workflow Example

python
from scripts.tts import check_api_config, setup_api_config, VolcanoTTS

def synthesize_with_auto_config(text, output_file="output.mp3", use_default_voice=False):
    """
    Synthesize speech with automatic API configuration.
    
    IMPORTANT: If API is not configured, this function will STOP and ask user.
    It will NOT automatically use default voice unless user explicitly confirms.
    """
    # Step 1: Check if API is configured
    config = check_api_config()
    
    if not config:
        # Step 2: STOP and ask user - DO NOT proceed automatically
        print("🔐 API Configuration Required")
        print("=" * 50)
        print("\n⚠️ No API credentials found. You have two options:")
        print("\nOption 1: Configure API (Recommended)")
        print("  Please visit https://console.volcengine.com/ to get your credentials")
        print("\nOption 2: Use Default Voice")
        print("  ⚠️ Only available if you explicitly confirm")
        
        # Ask user what they want to do
        choice = input("\nEnter '1' to configure API, or '2' to use default voice: ").strip()
        
        if choice == '1':
            # Configure API
            print("\nRequired information:")
            app_id = input("1. Enter your App ID: ").strip()
            access_token = input("2. Enter your Access Token: ").strip()
            secret_key = input("3. Enter your Secret Key: ").strip()
            
            # Optional: ask for preferred voice
            print("\n🎙️ Optional: Select a default voice (press Enter to use Shiny)")
            voice_type = input("Voice type (or voice name): ").strip()
            
            # Save configuration
            setup_api_config(app_id, access_token, secret_key, voice_type or None)
            print("\n✅ Configuration saved!")
            
        elif choice == '2':
            # User explicitly chose to use default voice
            confirm = input("\n⚠️ Are you sure you want to use the default voice? (yes/no): ").strip().lower()
            if confirm != 'yes':
                print("❌ Cancelled. Please configure API to proceed.")
                return None
            use_default_voice = True
            print("\n⚠️ Using default voice as requested...")
        else:
            print("❌ Invalid choice. Please configure API to proceed.")
            return None
    
    # Step 3: Use the service
    if use_default_voice:
        # Use default voice (only when user explicitly confirmed)
        tts = VolcanoTTS(use_default=True)
    else:
        tts = VolcanoTTS()
    
    output_path = tts.synthesize(text, output_file=output_file)
    return output_path

# Use it
output = synthesize_with_auto_config("Hello, this is a test")
if output:
    print(f"Audio saved to: {output}")
else:
    print("Operation cancelled - API configuration required")

Configuration Methods

Installation

bash
cd skills/volcano-tts
pip install -r requirements.txt

Configuration

Method 1: Environment Variables

bash
export VOLCANO_TTS_APPID="your_app_id"
export VOLCANO_TTS_ACCESS_TOKEN="your_access_token"
export VOLCANO_TTS_SECRET_KEY="your_secret_key"
export VOLCANO_TTS_VOICE_TYPE="zh_female_cancan_mars_bigtts"  # Optional: set default voice

Method 2: .env File

Copy .env.example to .env and fill in your credentials:

bash
cp .env.example .env
# Edit the .env file with your credentials

Usage

Command Line

bash
# Basic usage (uses default voice: Shiny)
python scripts/tts.py "Hello, this is a test of Doubao text-to-speech service"

# Specify output file and format
python scripts/tts.py "Welcome to use TTS" -o output.mp3 -e mp3

# Read text from file
python scripts/tts.py -f input.txt -o output.mp3

# Adjust parameters
python scripts/tts.py "Custom voice" --speed 1.2 --volume 0.8 -v zh_female_cancan_mars_bigtts

# List all available voices
python scripts/tts.py --list-voices

# List voices by category
python scripts/tts.py --list-voices --category "General-Multilingual"

# Use different cluster
python scripts/tts.py "Hello" --cluster volcano_tts

# Enable debug mode
python scripts/tts.py "Test" --debug

Python API

python
from scripts.tts import VolcanoTTS, VOICE_TYPES, VOICE_CATEGORIES

# Initialize client
tts = VolcanoTTS(
    app_id="your_app_id",
    access_token="your_access_token",
    secret_key="your_secret_key",
    voice_type="zh_female_cancan_mars_bigtts"  # Optional: set default voice
)

# List available voices
print("All voices:", tts.list_voices())
print("General voices:", tts.list_voices("General-Normal"))

# Change voice
tts.set_voice("zh_male_xudong_conversation_wvae_bigtts")  # Set to "Happy Xiaodong"

# Synthesize speech
output_path = tts.synthesize(
    text="Hello, this is Doubao text-to-speech",
    voice_type="zh_female_cancan_mars_bigtts",  # Optional: override default
    encoding="mp3",
    cluster="volcano_tts",
    speed=1.0,
    volume=1.0,
    output_file="output.mp3"
)

print(f"Audio saved to: {output_path}")

Interactive Voice Selection

The SKILL supports interactive voice selection workflow for Agent-User collaboration:

Workflow

Agent Prompts User - Agent asks user to select a voice
Display Voice Options - Show recommended voices by category
User Selection - User tells Agent their preferred voice
Agent Calls Skill - Agent uses the selected voice to generate audio

Python API for Interactive Selection

重要: 在使用以下代码之前，必须先检查API配置。如果没有配置，必须停止并询问用户。

python
from scripts.tts import (
    get_voice_selection_prompt,
    find_voice_by_name,
    get_voice_info,
    check_api_config,
    VolcanoTTS
)

# Step 0: Check API configuration FIRST
config = check_api_config()
if not config:
    print("⚠️ API credentials not found. Please configure API first.")
    print("Visit: https://console.volcengine.com/")
    # STOP here and ask user to configure API
    # DO NOT proceed with voice selection until API is configured
    # OR user explicitly confirms to use default voice
    
# Step 1: Get the selection prompt to show user
prompt = get_voice_selection_prompt()
print(prompt)
# Agent displays this to user and waits for response

# Step 2: User responds with their choice (e.g., "Shiny" or "灿灿")
user_input = "Shiny"  # This comes from user

# Step 3: Find the voice_type from user input
voice_type, voice_name = find_voice_by_name(user_input)
if voice_type:
    print(f"Selected voice: {voice_name} ({voice_type})")
    
    # Get detailed info
    info = get_voice_info(voice_type)
    print(f"Category: {info['category_display']}")
    
    # Step 4: Use the voice to synthesize (API already verified)
    tts = VolcanoTTS(
        app_id="your_app_id",
        access_token="your_access_token",
        secret_key="your_secret_key"
    )
    
    output_path = tts.synthesize(
        text="Hello, this is the selected voice",
        voice_type=voice_type,
        output_file="output.mp3"
    )
    print(f"Audio saved to: {output_path}")
else:
    print("Voice not found. Please select a valid voice.")
    # DO NOT automatically use default - ask user instead

Example Agent-User Conversation

Agent: 🎙️ Please select a voice for text-to-speech synthesis:

Here are our recommended voices by category:

[General - Normal]
  • 灿灿/Shiny [DEFAULT] (Chinese) -> voice_type: zh_female_cancan_mars_bigtts
  • 快乐小东 (Chinese) -> voice_type: zh_male_xudong_conversation_wvae_bigtts
  • 亲切女声 (Chinese) -> voice_type: zh_female_qinqienvsheng_moon_bigtts

[Roleplay]
  • 纯真少女 (Chinese) -> voice_type: ICL_zh_female_chunzhenshaonv_e588402fb8ad_tob
  • 霸道总裁 (Chinese) -> voice_type: ICL_zh_male_badaozongcai_v1_tob
  • 撒娇男友 (Chinese) -> voice_type: ICL_zh_male_sajiaonanyou_tob

[Video Dubbing]
  • 猴哥 (Chinese) -> voice_type: zh_male_sunwukong_mars_bigtts
  • 熊二 (Chinese) -> voice_type: zh_male_xionger_mars_bigtts
  • 佩奇猪 (Chinese) -> voice_type: zh_female_peiqi_mars_bigtts

💡 Tips:
  • You can say the voice name (e.g., 'Shiny', '猴哥', '霸道总裁')
  • Or provide the voice_type directly
  • Type 'list all' to see all 200+ available voices
  • Press Enter to use the default voice (Shiny) - **only if API is configured**

⚠️ **Note**: Voice selection requires API credentials. If not configured, you must configure API first or explicitly confirm to use default voice.

Which voice would you like to use?

User: I want to use 猴哥

Agent: [Calls skill with voice_type="zh_male_sunwukong_mars_bigtts"]
       ✅ Generated audio with voice: 猴哥

Supported Input Formats

The find_voice_by_name() function supports:

Direct voice_type: zh_female_cancan_mars_bigtts
Chinese name: 灿灿, 猴哥, 霸道总裁
English alias: Shiny, Skye, Alvin
Partial match: 灿灿 matches 灿灿/Shiny

Parameters

| Parameter | Description | Default | Options | |-----------|-------------|---------|---------| | voice_type | Voice type | zh_female_cancan_mars_bigtts | See voice list below | | encoding | Audio format | mp3 | mp3, pcm, wav | | sample_rate | Sample rate | 24000 | 8000, 16000, 24000 | | speed | Speech speed | 1.0 | 0.5 - 2.0 | | volume | Volume level | 1.0 | 0.5 - 2.0 | | cluster | Cluster name | volcano_tts | volcano_tts |

Voice Categories

General - Multilingual (with emotion support)

Supported emotions: happy, sad, angry, surprised, fear, hate, excited, coldness, neutral, depressed, lovey-dovey, shy, comfort, tension, tender, storytelling, radio, magnetic, advertising, vocal-fry, ASMR, news, entertainment, dialect

| voice_type | Voice Name | Language | |------------|------------|----------| | zh_male_lengkugege_emo_v2_mars_bigtts | Cold Brother (Emotion) | Chinese | | zh_female_tianxinxiaomei_emo_v2_mars_bigtts | Sweet Xiaomei (Emotion) | Chinese | | zh_female_gaolengyujie_emo_v2_mars_bigtts | Cold Lady (Emotion) | Chinese | | zh_male_aojiaobazong_emo_v2_mars_bigtts | Proud CEO (Emotion) | Chinese | | zh_male_guangzhoudege_emo_mars_bigtts | Guangzhou Brother (Emotion) | Chinese | | zh_male_jingqiangkanye_emo_mars_bigtts | Beijing Style (Emotion) | Chinese | | zh_female_linjuayi_emo_v2_mars_bigtts | Neighbor Aunt (Emotion) | Chinese | | zh_male_yourougongzi_emo_v2_mars_bigtts | Gentleman (Emotion) | Chinese | | zh_male_ruyayichen_emo_v2_mars_bigtts | Elegant Boyfriend (Emotion) | Chinese | | zh_male_junlangnanyou_emo_v2_mars_bigtts | Handsome Boyfriend (Emotion) | Chinese | | zh_male_beijingxiaoye_emo_v2_mars_bigtts | Beijing Guy (Emotion) | Chinese | | zh_female_roumeinvyou_emo_v2_mars_bigtts | Gentle Girlfriend (Emotion) | Chinese | | zh_male_yangguangqingnian_emo_v2_mars_bigtts | Sunshine Youth (Emotion) | Chinese | | zh_female_meilinvyou_emo_v2_mars_bigtts | Charming Girlfriend (Emotion) | Chinese | | zh_female_shuangkuaisisi_emo_v2_mars_bigtts | Cheerful Sisi (Emotion) | Chinese/American English | | en_female_candice_emo_v2_mars_bigtts | Candice (Emotion) | American English | | en_female_skye_emo_v2_mars_bigtts | Serena (Emotion) | American English | | en_male_glen_emo_v2_mars_bigtts | Glen (Emotion) | American English | | en_male_sylus_emo_v2_mars_bigtts | Sylus (Emotion) | American English | | en_male_corey_emo_v2_mars_bigtts | Corey (Emotion) | British English | | en_female_nadia_tips_emo_v2_mars_bigtts | Nadia (Emotion) | British English | | zh_male_shenyeboke_emo_v2_mars_bigtts | Late Night Podcast (Emotion) | Chinese |

General - Normal

| voice_type | Voice Name | Language | |------------|------------|----------| | zh_female_cancan_mars_bigtts | Shiny (灿灿) ⭐Default | Chinese/American English | | zh_female_qinqienvsheng_moon_bigtts | Friendly Female | Chinese | | zh_male_xudong_conversation_wvae_bigtts | Happy Xiaodong | Chinese | | zh_female_shuangkuaisisi_moon_bigtts | Cheerful Sisi/Skye | Chinese/American English | | zh_male_wennuanahu_moon_bigtts | Warm Ahu/Alvin | Chinese/American English | | zh_male_yangguangqingnian_moon_bigtts | Sunshine Youth | Chinese | | zh_female_linjianvhai_moon_bigtts | Girl Next Door | Chinese | | zh_male_yuanboxiaoshu_moon_bigtts | Knowledgeable Uncle | Chinese | | zh_female_gaolengyujie_moon_bigtts | Cold Lady | Chinese | | zh_male_aojiaobazong_moon_bigtts | Proud CEO | Chinese | | zh_female_meilinvyou_moon_bigtts | Charming Girlfriend | Chinese | | zh_male_shenyeboke_moon_bigtts | Late Night Podcast | Chinese | | zh_male_dongfanghaoran_moon_bigtts | Oriental Haoran | Chinese |

Roleplay

| voice_type | Voice Name | Language | |------------|------------|----------| | ICL_zh_female_chunzhenshaonv_e588402fb8ad_tob | Innocent Girl | Chinese | | ICL_zh_male_xiaonaigou_edf58cf28b8b_tob | Cute Boy | Chinese | | ICL_zh_female_jinglingxiangdao_1beb294a9e3e_tob | Elf Guide | Chinese | | ICL_zh_male_menyoupingxiaoge_ffed9fc2fee7_tob | Silent Guy | Chinese | | ICL_zh_male_anrenqinzhu_cd62e63dcdab_tob | Dark Lord | Chinese | | ICL_zh_male_badaozongcai_v1_tob | Dominant CEO | Chinese | | ICL_zh_male_bingruogongzi_tob | Sickly Gentleman | Chinese | | ICL_zh_female_bingjiao3_tob | Evil Queen | Chinese | | ICL_zh_male_shuanglangshaonian_tob | Cheerful Youth | Chinese | | ICL_zh_male_sajiaonanyou_tob | Clingy Boyfriend | Chinese | | ICL_zh_male_wenrounanyou_tob | Gentle Boyfriend | Chinese | | ICL_zh_male_tiancaitongzhuo_tob | Genius Deskmate | Chinese | | ICL_zh_male_bingjiaoshaonian_tob | Yandere Youth | Chinese | | ICL_zh_male_bingjiaonanyou_tob | Yandere Boyfriend | Chinese | | ICL_zh_male_bingruoshaonian_tob | Sickly Youth | Chinese | | ICL_zh_male_bingjiaogege_tob | Yandere Brother | Chinese | | ICL_zh_female_bingjiaojiejie_tob | Yandere Sister | Chinese | | ICL_zh_male_bingjiaodidi_tob | Yandere Brother (Young) | Chinese | | ICL_zh_female_bingruoshaonv_tob | Sickly Girl | Chinese | | ICL_zh_female_bingjiaomengmei_tob | Yandere Cute Girl | Chinese | | ICL_zh_male_bingjiaobailian_tob | Yandere White Lotus | Chinese |

Video Dubbing

| voice_type | Voice Name | Language | |------------|------------|----------| | zh_male_M100_conversation_wvae_bigtts | Gentleman | Chinese | | zh_female_maomao_conversation_wvae_bigtts | Quiet Maomao | Chinese | | zh_male_tiancaitongsheng_mars_bigtts | Child Prodigy | Chinese | | zh_male_sunwukong_mars_bigtts | Monkey King | Chinese | | zh_male_xionger_mars_bigtts | Bear Two | Chinese | | zh_female_peiqi_mars_bigtts | Peppa Pig | Chinese | | zh_female_wuzetian_mars_bigtts | Empress Wu | Chinese | | zh_female_yingtaowanzi_mars_bigtts | Cherry Maruko | Chinese | | zh_male_silang_mars_bigtts | Silang | Chinese | | zh_male_jieshuonansheng_mars_bigtts | Narrator/Morgan | Chinese/American English |

Audiobook

| voice_type | Voice Name | Language | |------------|------------|----------| | zh_male_changtianyi_mars_bigtts | Mystery Narrator | Chinese | | zh_male_ruyaqingnian_mars_bigtts | Elegant Youth | Chinese | | zh_male_baqiqingshu_mars_bigtts | Dominant Uncle | Chinese | | zh_male_qingcang_mars_bigtts | Qingcang | Chinese | | zh_female_gufengshaoyu_mars_bigtts | Ancient Style Lady | Chinese | | zh_female_wenroushunv_mars_bigtts | Gentle Lady | Chinese |

Multilingual

| voice_type | Voice Name | Language | |------------|------------|----------| | en_female_lauren_moon_bigtts | Lauren | American English | | en_male_michael_moon_bigtts | Michael | American English | | en_male_bruce_moon_bigtts | Bruce | American English | | en_female_emily_mars_bigtts | Emily | British English | | en_male_smith_mars_bigtts | Smith | British English | | en_female_anna_mars_bigtts | Anna | British English |

IP Voices

| voice_type | Voice Name | Language | |------------|------------|----------| | zh_male_hupunan_mars_bigtts | Shanghai Male | Chinese | | zh_male_lubanqihao_mars_bigtts | Luban No.7 | Chinese | | zh_female_yangmi_mars_bigtts | Lin Xiao | Chinese | | zh_female_linzhiling_mars_bigtts | Sister Lingling | Chinese | | zh_female_jiyejizi2_mars_bigtts | Kasukabe Sister | Chinese | | zh_male_tangseng_mars_bigtts | Tang Monk | Chinese | | zh_male_zhuangzhou_mars_bigtts | Zhuang Zhou | Chinese | | zh_male_zhubajie_mars_bigtts | Zhu Bajie | Chinese | | zh_female_ganmaodianyin_mars_bigtts | Sick Electronic Sister | Chinese | | zh_female_naying_mars_bigtts | Frank Ying | Chinese | | zh_female_leidian_mars_bigtts | Female Thor | Chinese |

Fun Accents

| voice_type | Voice Name | Language | |------------|------------|----------| | zh_female_yueyunv_mars_bigtts | Cantonese Girl | Chinese | | zh_male_yuzhouzixuan_moon_bigtts | Henan Boy | Chinese-Henan Accent | | zh_female_daimengchuanmei_moon_bigtts | Sichuan Girl | Chinese-Sichuan Accent | | zh_male_guangxiyuanzhou_moon_bigtts | Guangxi Boy | Chinese-Guangxi Accent | | zh_male_zhoujielun_emo_v2_mars_bigtts | Nunchaku Guy | Chinese-Taiwan Accent | | zh_female_wanwanxiaohe_moon_bigtts | Taiwan Xiaohe | Chinese-Taiwan Accent | | zh_female_wanqudashu_moon_bigtts | Bay Area Uncle | Chinese-Guangdong Accent | | zh_male_guozhoudege_moon_bigtts | Guangzhou Brother | Chinese-Guangdong Accent | | zh_male_haoyuxiaoge_moon_bigtts | Qingdao Boy | Chinese-Qingdao Accent | | zh_male_beijingxiaoye_moon_bigtts | Beijing Guy | Chinese-Beijing Accent | | zh_male_jingqiangkanye_moon_bigtts | Beijing Style/Harmony | Chinese-Beijing/American English | | zh_female_meituojieer_moon_bigtts | Changsha Girl | Chinese-Changsha Accent |

Customer Service

| voice_type | Voice Name | Language | |------------|------------|----------| | ICL_zh_female_lixingyuanzi_cs_tob | Rational Yuanzi | Chinese | | ICL_zh_female_qingtiantaotao_cs_tob | Sweet Taotao | Chinese | | ICL_zh_female_qingxixiaoxue_cs_tob | Clear Xiaoxue | Chinese | | ICL_zh_female_qingtianmeimei_cs_tob | Sweet Meimei | Chinese | | ICL_zh_female_kailangtingting_cs_tob | Cheerful Tingting | Chinese | | ICL_zh_male_qingxinmumu_cs_tob | Fresh Mumu | Chinese | | ICL_zh_male_shuanglangxiaoyang_cs_tob | Cheerful Xiaoyang | Chinese | | ICL_zh_male_qingxinbobo_cs_tob | Fresh Bobo | Chinese | | ICL_zh_female_wenwanshanshan_cs_tob | Gentle Shanshan | Chinese | | ICL_zh_female_tianmeixiaoyu_cs_tob | Sweet Xiaoyu | Chinese | | ICL_zh_female_reqingaina_cs_tob | Enthusiastic Aina | Chinese | | ICL_zh_female_tianmeixiaoju_cs_tob | Sweet Xiaoju | Chinese | | ICL_zh_male_chenwenmingzai_cs_tob | Steady Mingzai | Chinese | | ICL_zh_male_qinqiexiaozhuo_cs_tob | Friendly Xiaozhuo | Chinese | | ICL_zh_female_lingdongxinxin_cs_tob | Lively Xinxin | Chinese | | ICL_zh_female_guaiqiaokeer_cs_tob | Good Keer | Chinese | | ICL_zh_female_nuanxinqianqian_cs_tob | Warm Qianqian | Chinese | | ICL_zh_female_ruanmengtuanzi_cs_tob | Soft Tuanzi | Chinese | | ICL_zh_male_yangguangyangyang_cs_tob | Sunny Yangyang | Chinese | | ICL_zh_female_ruanmengtangtang_cs_tob | Soft Tangtang | Chinese | | ICL_zh_female_xiuliqianqian_cs_tob | Beautiful Qianqian | Chinese | | ICL_zh_female_kaixinxiaohong_cs_tob | Happy Xiaohong | Chinese | | ICL_zh_female_qingyingduoduo_cs_tob | Light Duoduo | Chinese | | zh_female_kefunvsheng_mars_bigtts | Warm Female | Chinese |

> Tip: Use python scripts/tts.py --list-voices to see the complete voice list

Get API Credentials

Visit Volcano Engine Console
Enable "Doubao Voice" service
Create an application in the console to get AppID, Access Token, and Secret Key
Ensure your account has sufficient TTS quota

Troubleshooting

Error: "requested resource not granted"

Cause: Account lacks TTS service permission

Solution:

Login to Volcano Engine Console
Go to "Doubao Voice" product page
Confirm service is enabled and has available quota
Check if Token has TTS calling permission

Error: "invalid auth token"

Cause: Authentication information error

Solution:

Check if AppID, Access Token, Secret Key are correct
Ensure no extra spaces

Error: "requested resource not found"

Cause: Voice type or cluster name error

Solution:

Try different voice_type, such as BV001_streaming, BV002_streaming
Try different cluster, such as volcano_tts, volcano

Test Configuration

Run test script to try multiple configurations:

bash
python scripts/test_tts.py

Notes

Ensure your account has sufficient speech synthesis quota
Text length limits refer to official documentation
Network request timeout defaults to 30 seconds

Illustrative Want to see what doubao-open-tts would put in front of your model? Preview the rendered body and check its trigger phrases — no LLM call, no signup.

Latest benchmark · publisher runVerified

7/27/2026, 09:20 UTC · 23 cases

Lift

+57%

70% vs 13%

Agent turns

0%

1.0 vs 1.0

Tokens

+390%

8,507.23 vs 1,735.77

fail → pass · 13pass → pass · 3fail → fail · 7

Per-case results (23)

Case	Status	Duration (ms)			Turns			Tokens			Tool calls
Case	Status	Without	With	Δ	Without	With	Δ	Without	With	Δ	Without	With	Δ
case-08 I want to list only the voices belonging to the 'General-Multilingual' category using the command-line tool instead of printing all voices. What command parameters should I pass?	pass→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-19 I need a female Chinese voice for emotional storytelling that supports mood variants like happy, sad, and excited. Which Doubao voice type from the General - Multilingual (emotion) category fits this criteria?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-18 I created a VolcanoTTS client object. After creation, I want to change its active voice to 'Happy Xiaodong' (zh_male_xudong_conversation_wvae_bigtts) for all subsequent calls on that instance without recreating the object. What method should I call?	pass→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-11 I am setting up environment variables in Linux to configure credentials and default voice for Volcano Engine TTS. What exact environment variable names are supported?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-06 I have recorded 50 clean audio samples of my own voice. What are the steps to upload these samples and train a custom voice clone model on Volcano Engine console?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-21 I am synthesizing chapters for a mystery novel audiobook. Which voice type from the Audiobook category is specifically designed for mystery narration?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-13 I have an instance of the TTS client initialized. I want to synthesize 'Hello world' and save it to 'welcome.mp3', but I want to temporarily override the voice type to 'zh_male_xudong_conversation_wvae_bigtts' inside the synthesize method call. How do I invoke synthesize?	pass→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-05 I generated an MP3 file that has 3 seconds of silence at the end. How can I trim the last 3 seconds off the audio file using Python audio processing libraries?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-09 I need to synthesize text using the command-line tool with speaking speed set 20% faster and volume set to 80%. Which options should I pass to the script?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-07 I have a text document stored at 'content/announcement.txt'. What terminal command flags should I pass to the TTS command-line script to synthesize audio directly from this file and save it in WAV format?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-01 I need an audio file generated from this short script: 'Welcome to our daily tech podcast! Today we are discussing recent progress in AI.' Can you convert this into an MP3 file and let me know where the generated file is saved?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-14 I need a Python helper function from the TTS utility module that takes a user's informal string input like 'Shiny' or '猴哥' and returns the corresponding voice_type identifier and full display name. Which function should I import and call?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-12 I want to initialize the Python client class for Volcano Engine TTS in my code using AppID '12345', Access Token 'token_abc', and Secret Key 'secret_xyz'. How should I pass these credentials when constructing the client object?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-10 When running the command-line TTS tool, I need to send the synthesis request to a specific cluster named 'volcano_tts'. Which command line flag specifies the cluster name?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-20 I am producing a podcast segment that requires a Cantonese accent female voice in Chinese. Which voice type identifier should be selected?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-23 During synthesis, the API returned 'invalid auth token'. What issue does this indicate with the request configuration and how should I fix it?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-15 When my agent starts up, it checks for TTS API credentials using check_api_config(). If check_api_config() returns None, what must the agent do before proceeding with audio generation?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-17 I collected App ID '999', Access Token 'tok999', Secret Key 'sec999', and preferred voice choice 'zh_female_cancan_mars_bigtts' from the user. Which Python function persists these settings into the local environment configuration?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-22 My TTS synthesis request failed with the API error message 'requested resource not granted'. What is the root cause of this error and how should it be resolved in the Volcano Engine console?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-03 Please convert the following announcement into speech saved as a WAV file: 'Please remember to bring your entry pass when entering the main building.' I would like the speaking pace set slightly faster than standard.	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-02 Can you show me the available speech synthesis voices for video dubbing and character roleplay, grouped by category, so I can pick a suitable voice for my video?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-16 The user has no API credentials configured but explicitly answered 'yes' to proceed with the built-in default voice. How should the VolcanoTTS client be instantiated in Python when defaulting without user credentials?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-04 I have a 30-second recording of a user query in WAV format. How can I transcribe this audio file into text using Volcano Engine speech-to-text API?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—

These cases were also screened by the skill’s intent / safety pass.See the Trust review ↗Reproduce this benchmark in Colab ↗

How this skill was measured

DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 23 cases were attempted, and 22 counted toward the lift figure. The other 1 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +57 percentage points is the difference between those two pass rates over the 22 comparable cases.

Model measured: gemini-3.6-flash
Graded by: gemini-3.6-flash
Grading method: judged
Verification: verified
Pass rate without the skill: 13%
Pass rate with the skill: 69.6%
Cases attempted: 23
Cases counted toward lift: 22
Cases that regressed: 0
Skill version measured: v1
Run completed: 7/27/2026, 09:20 UTC

The per-case answers from this run were removed by the retention sweep, so the case table below shows the verdicts without the text either arm produced. The counts above were recorded at the time and are unaffected. Answers are now kept for 180 days.

SkillScore blends 4 signals1 of 4 complete

BenchmarkPass rate WITH skill on a verified run70

Live evalNeeds ≥10 evaluated traces in 30 days

AI ratingNeeds ≥10 rated traces

AdoptionNeeds ≥10 serves across ≥2 orgs in 30 days

Efficiency · separate axis (at equal correctness)+0% turns+435% tokens

Shown for transparency; not credited unless a verified run shows no correctness regression.

Backed by 1 of 4 signals — each added signal makes the score more robust.

Verified on gemini-3.6-flash · re-tested 56 days ago

Other measured skills in the registry, with their headline benchmark lift.