Volcano Engine Real-time Audio and Video

https://www.volcengine.com/docs/6348/2123348
Admin
Volcano Engine Real-time Audio and Video is a platform providing SDKs and APIs for implementing...
Tokens:32,326
Snippets:176
Trust Score:10
Update:9 months ago
Show doc for...
Tokens:
Raw
### StartVoiceChat API Request Example (JSON)

Source: https://www.volcengine.com/docs/6348/2123348/index

This JSON object demonstrates a complete request payload for the StartVoiceChat API. It includes configurations for ASR, LLM, TTS, and agent behavior, such as target user, welcome message, and burst object settings.

```json
{
  "AppId": "68e*****477",
  "RoomId": "ChatRoom01",
  "TaskId": "ChatTask01", 
  "Config": {
    "ASRConfig": {
      "Provider": "volcano",
      "ProviderParams": {
        "Mode": "bigmodel",
        "StreamMode": 0
      },
      "VADConfig": {
        "SilenceTime": 600,
        "AIVAD": true
      },
      "InterruptConfig": {}
    },
    "LLMConfig": {
      "Mode": "ArkV3",      "ModelName": "doubao-seed-1-8-251215",      "SystemMessages": [
        "##人设\n你是一个全能的超级助手，具备强大的知识库、情感理解能力和解决问题的能力。你的目标是高效、专业、友好地帮助用户完成各类任务，包括但不限于日常生活、工作安排、信息检索、学习辅导、创意写作、语言翻译和技术支持等；\n\n##约束\n始终主动、礼貌、有条理；\n回答准确但不冗长，必要时可提供简洁总结+详细解释；\n不清楚的任务会主动澄清，不假设、不误导。"
      ],
      "HistoryLength": 10,
      "ThinkingType": "disabled",
      "VisionConfig": {
        "SnapshotConfig": {},
        "StorageConfig": {
          "TosConfig": {}
        }
      }
    },
    "TTSConfig": {
      "Provider": "volcano_bidirection",
      "ProviderParams": {
        "Credential": {
          "ResourceId": "seed-tts-2.0"
        },
        "VolcanoTTSParameters": "{\"req_params\":{\"speaker\":\"ICL_zh_female_wuxi_tob\",\"audio_params\":{\"speech_rate\":0}}}"
      }
    },
    "SubtitleConfig": {},
    "InterruptMode": 0,
    "FunctionCallingConfig": {},
    "WebSearchAgentConfig": {},
    "MemoryConfig": {},
    "MusicAgentConfig": {}
  },
  "AgentConfig": {
    "TargetUserId": ["Huoshan01"],
    "UserId": "ChatBot01",
    "WelcomeMessage": "我是你的AI助手，有什么需要我为您效劳的吗？",
    "Burst": {
      "Enable": false,
      "BufferSize": 0,
      "Interval": 0
    },
    "VoicePrint": {
      "MetaList": null,
      "VoicePrintList": null
    }
  }
}
```

--------------------------------

### Third-Party LLM/Agent Configuration - System and User Prompts

Source: https://www.volcengine.com/docs/6348/2123348/index

Defines the system-level instructions and user-provided prompts to guide the behavior and responses of third-party models. SystemMessages set the AI's persona and rules, while UserPrompts enhance response quality.

```javascript
{
  "SystemMessages": ["You are a helpful assistant."],
  "UserPrompts": [
    { "Role": "user", "Content": "What is the weather like today?" },
    { "Role": "assistant", "Content": "The weather is sunny." }
  ]
}
```

--------------------------------

### StartVoiceChat API Response Example (JSON)

Source: https://www.volcengine.com/docs/6348/2123348/index

This JSON object represents a successful response from the StartVoiceChat API. It includes metadata about the request and a 'Result' field indicating success.

```json
{
    "Result": "ok",
    "ResponseMetadata": {
        "RequestId": "202512****10420",
        "Action": "StartVoiceChat",
        "Version": "2025-06-01",
        "Service": "rtc",
        "Region": "cn-north-1"
    }
}
```

--------------------------------

### Coze Custom Parameters Configuration

Source: https://www.volcengine.com/docs/6348/2123348/index

Enables dynamic assignment of values to custom parameters defined in the Coze conversation flow's starting node. This is applicable for published APIs and single-agent ChatSDK modes.

```json
{
  "custom_param_1": "some_value",
  "custom_param_2": 123
}
```

--------------------------------

### Volcengine Agent Basic Configuration

Source: https://www.volcengine.com/docs/6348/2123348/index

Configure essential agent settings, including the target user ID, welcome message, agent ID, and callback settings for conversation state.

```json
{
  "TargetUserId": ["user1"],
  "WelcomeMessage": "Hello",
  "UserId": "BotName001",
  "EnableConversationStateCallback": false
}
```

--------------------------------

### Volcengine License User Configuration

Source: https://www.volcengine.com/docs/6348/2123348/index

Configure settings for license users, indicating whether a license is in use.

```json
{
  "UseLicense": true
}
```

--------------------------------

### Configure Voice Cloning (Non-Streaming) (JSON)

Source: https://www.volcengine.com/docs/6348/2123348/index

This configuration is for Volcano Engine's voice cloning service with non-streaming input and output. It simplifies setup by encapsulating common parameters and requires specifying the voice type, cluster ID, and model version.

```json
{
  "audio": {
    "voice_type": "S_U****ft1"
  },
  "app": {
    "cluster": "volcano_icl"
  }
}
```

--------------------------------

### Third-Party LLM/Agent Configuration - Generation Parameters

Source: https://www.volcengine.com/docs/6348/2123348/index

Optional parameters for controlling the output of third-party Large Models/Agents. MaxTokens limits output length, Temperature adjusts randomness, and TopP controls token diversity.

```javascript
{
  "ModelName": "name1",
  "MaxTokens": 1024,
  "Temperature": 0.1,
  "TopP": 0.3
}
```

--------------------------------

### StartVoiceChat Configuration

Source: https://www.volcengine.com/docs/6348/2123348/index

This section outlines the various configuration options available for the StartVoiceChat API, including ASR, VAD, and TTS settings.

```APIDOC
## StartVoiceChat API

### Description
This API enables real-time voice chat with AI models, supporting various configurations for speech recognition, voice activity detection, and text-to-speech.

### Method
POST

### Endpoint
/websites/volcengine_6348_2123348

### Parameters
#### Request Body
- **context** (String) - Optional - Specifies hotwords to improve ASR accuracy. Format: JSON string. Max length: 200 tokens.
- **context_history_length** (Integer) - Optional - Number of recent conversation turns to include for context. Range: [0, 20]. Default: 0.
- **火山流式语音识别（小模型）** (Object) - Volcanic streaming ASR model configuration.
  - **Mode** (String) - Required - Model type. Fixed value: 'smallmodel'.
  - **Cluster** (String) - Required - Cluster ID for the ASR model. Example: 'volcengine_streaming_common'.
- **自定义语音识别** (Object) - Custom ASR model configuration.
  - **URL** (String) - Required - Gateway service access point URL. Format: 'wss://ai-gateway.vei.volces.com/v1/realtime?model=<ASR 调用名称>'.
  - **APIKey** (String) - Required - Gateway access key. Example: 'sk-xxxxxx'.
  - **ExtraData** (JSONMap) - Optional - Custom parameters passed to the ASR service.
- **VADConfig** (Object) - Optional - Voice Activity Detection configuration.
  - **SilenceTime** (Integer) - Optional - Silence duration for sentence end detection. Range: [500, 3000) ms. Default: 600.
  - **AIVAD** (Boolean) - Optional - Enable AI-semantic sentence segmentation. Default: false.
- **VolumeGain** (Float) - Optional - Volume gain value. Default: 1.0, Recommended: 0.3.
- **InterruptConfig** (Object) - Optional - Speech interruption configuration.
  - **InterruptSpeechDuration** (Integer) - Optional - Automatic interruption threshold based on speech duration. Range: 0 or [200, 3000] ms. Default: 0.
  - **InterruptKeywords** (String[]) - Optional - List of keywords to trigger interruption.
- **TurnDetectionMode** (Integer) - Optional - Whether to automatically trigger a new conversation round. 0: Auto-trigger (default), 1: Manual trigger.
- **TTSConfig** (Object) - Required - Text-to-Speech configuration.
  - **Provider** (String) - Required - TTS service provider. Options: 'volcano_bidirection', 'volcano', 'minimax', 'ai_gateway'.
  - **ProviderParams** (Object) - Required - Parameters for the selected TTS provider.
    - **火山语音合成大模型（流式输入流式输出_参数透传）**:
      - **Credential** (Object) - Optional - TTS model version credential.
        - **ResourceId** (String) - Optional - TTS model version. Options: 'seed-tts-1.0' or 'volc.service_type.10029' (default), 'seed-tts-2.0'.
      - **VolcanoTTSParameters** (String) - Required - JSON string for Volcanic TTS native API parameters. Must include 'req_params.speaker'.
    - **火山语音合成大模型（流式输入流式输出_参数直传）**:
      - **ResourceId** (String) - Optional - TTS model version. Options: 'seed-tts-1.0' or 'volc.service_type.10029' (default), 'seed-tts-2.0'.
      - **audio** (Object) - Required - Audio configuration for Volcanic TTS.
        - **voice_type** (String) - Required - Voice type, must match ResourceId. Example: 'zh_female_linjianvhai_moon_bigtts'.
        - **speech_rate** (Integer) - Optional - Speech rate. Range: [-50, 100]. Default: 0.

### Request Example
```json
{
  "context": "{\"hotwords\": [{\"word\": \"CO2\"},{\"word\": \"雨伞\"},{\"word\": \"鱼\"}]}",
  "context_history_length": 2,
  "火山流式语音识别（小模型）": {
    "Mode": "smallmodel",
    "Cluster": "volcengine_streaming_common"
  },
  "TTSConfig": {
    "Provider": "volcano_bidirection",
    "ProviderParams": {
      "火山语音合成大模型（流式输入流式输出_参数透传）": {
        "VolcanoTTSParameters": "{\"req_params\":{\"speaker\":\"zh_female_linjianvhai_moon_bigtts\", \"audio_params\": { \"speech_rate\": 100 }}}"
      }
    }
  }
}
```

### Response
#### Success Response (200)
- **Response structure will vary based on the streaming nature of the service and specific configurations.**

#### Response Example
```json
{
  "text": "Received your request.",
  "audio_stream": "<audio data>"
}
```
```

--------------------------------

### Configure TTS Parameters (Volcano Bidirectional - Direct)

Source: https://www.volcengine.com/docs/6348/2123348/index

Configures parameters for Volcengine's bidirectional streaming TTS service directly, offering more flexibility than passing a JSON string. Includes specifying the voice type and adjusting speech rate.

```json
{
  "audio": {
    "voice_type": "zh_female_linjianvhai_moon_bigtts",
    "speech_rate": 0
  }
}
```

--------------------------------

### Music Agent Configuration

Source: https://www.volcengine.com/docs/6348/2123348/index

Configure settings for the AI music agent.

```APIDOC
## Music Agent Configuration

### Description
Configuration for the AI music agent, allowing the AI to play music or control music playback based on user commands.

### Parameters
#### Request Body
- **Enable** (Boolean) - Optional - Whether to enable the AI music agent. Defaults to `false`. If enabled, system messages in `LLMConfig.SystemMessages` should be updated to guide the model to use the `music_player` tool.
```

--------------------------------

### Volcengine Agent Server Callback Configuration

Source: https://www.volcengine.com/docs/6348/2123348/index

Configure server-side callback for agent state changes, specifying the server URL and a signature for verifying request authenticity.

```json
{
  "ServerMessageURLForRTS": "https://example-domain.com/vertc/callback",
  "ServerMessageSignatureForRTS": "b46ab5f1d8ad6a"
}
```

--------------------------------

### Configure Voice Cloning (Direct Parameter) (JSON)

Source: https://www.volcengine.com/docs/6348/2123348/index

This snippet demonstrates the direct parameter configuration for Volcano Engine's voice cloning service (streaming input, streaming output). It simplifies integration by encapsulating common parameters, such as voice type and speech rate, along with the resource ID for model version selection.

```json
{
  "audio": {
    "voice_type": "S_U****ft1",
    "speech_rate": 50
  },
  "ResourceId": "seed-icl-1.0"
}
```

--------------------------------

### MCP Object for Knowledge Base Integration

Source: https://www.volcengine.com/docs/6348/2123348/index

Configuration for integrating with external tool services like knowledge bases via MCP (Multi-Channel Platform). It requires a URL, a name, and optionally comfort words and a list of interested tools.

```json
{
  "URL": "https://knowledge-mcp-server.com/test/123",
  "Name": "knowledge",
  "ComfortWords": "正在处理中...",
  "InterestedTools": ["search_knowledge"]
}
```

--------------------------------

### Third-Party LLM/Agent Configuration - Basic

Source: https://www.volcengine.com/docs/6348/2123348/index

Configuration for integrating external Large Models or Agents. Mode must be 'CustomLLM', and Url points to the service endpoint. APIKey provides Bearer Token authentication.

```javascript
{
  "Mode": "CustomLLM",
  "Url": "https://test.com/path/to/app",
  "APIKey": "pat*****123231"
}
```

--------------------------------

### Volcengine AI Music Agent Configuration

Source: https://www.volcengine.com/docs/6348/2123348/index

Configure the AI music agent for Volcengine, enabling it and providing guidance on system messages for invoking the music_player tool.

```json
{
  "Enable": true
}
```

--------------------------------

### StartVoiceChat API

Source: https://www.volcengine.com/docs/6348/2123348/index

The StartVoiceChat API allows you to launch a voice and hearing-enabled audio-visual interactive intelligent agent, enabling natural, fluent, and human-like real-time conversations with human users.

```APIDOC
## POST /StartVoiceChat

### Description
Initiates a voice chat intelligent agent for real-time interactive conversations.

### Method
POST

### Endpoint
https://rtc.volcengineapi.com?Action=StartVoiceChat&Version=2025-06-01

### Parameters
#### Query Parameters
- **Action** (string) - Required - Example: StartVoiceChat
- **Version** (string) - Required - Example: 2025-06-01

#### Request Body
- **AppId** (string) - Required - The AppId for the audio-visual interactive intelligent agent. Must be from an audio-visual interactive agent application and consistent with the AppId used for generating the RTC authentication token.
- **RoomId** (string) - Required - The ID of the RTC room. Must be consistent with the RoomId used for generating the RTC authentication token.
- **TaskId** (string) - Required - A unique ID defined by you for the agent task. Must be used in subsequent UpdateVoiceChat or StopVoiceChat calls. Must be unique within the same AppId and RoomId combination.
- **Config** (object) - Required - Configuration for the agent interaction service, including ASR, TTS, LLM, captions, and Function Calling.
  - **ASRConfig** (object) - Required - Configuration for Automatic Speech Recognition (ASR).
    - **Provider** (string) - Required - The provider of the speech recognition service. Options: 'volcano' (for Volcano Engine's Doubao speech service) or 'ai_gateway' (for custom models via Volcano Edge LLM Gateway).
    - **ProviderParams** (object) - Required - Service configuration parameters. Varies based on the provider.
      - **Mode** (string) - Required (for Volcano Engine ASR) - Model type. Fixed value: 'bigmodel' for Volcano Engine's large model ASR.
      - **VolcanoASRParameters** (string) - Optional (for Volcano Engine ASR, parameter passthrough) - A JSON string for passing native API parameters to the Volcano Engine large model ASR service. Leave as '{}' if no custom parameters are needed.
      - **ApiResourceId** (string) - Optional (for Volcano Engine ASR) - Specifies the large model version (e.g., 'volc.bigasr.sauc.duration', 'volc.seedasr.sauc.duration').
      - **StreamMode** (integer) - Optional (for Volcano Engine ASR) - ASR mode: 0 (streaming input/output, faster), 1 (streaming input/non-streaming output, more accurate), 2 (recommended, bidirectional streaming optimized).
      - **ContextHistoryLength** (integer) - Optional (for Volcano Engine ASR) - Number of recent conversation turns to include for context (0-20). Defaults to 0.
      - **enable_nonstream** (boolean) - Optional (for Volcano Engine ASR, only with StreamMode 2) - Enables secondary recognition on detected full sentence segments when StreamMode is 2.

### Request Example
```json
{
  "AppId": "661*****3cf",
  "RoomId": "Room1",
  "TaskId": "task1",
  "Config": {
    "ASRConfig": {
      "Provider": "volcano",
      "ProviderParams": {
        "Mode": "bigmodel",
        "VolcanoASRParameters": "{\"request\":{\"vad_segment_duration\":1000,\"end_window_size\":500}}",
        "ApiResourceId": "volc.bigasr.sauc.duration",
        "StreamMode": 2,
        "enable_nonstream": true,
        "ContextHistoryLength": 5
      }
    }
  }
}
```

### Response
#### Success Response (200)
- **message** (string) - Indicates the success of the agent task initiation.

#### Response Example
```json
{
  "message": "success"
}
```

### Error Handling
- **400 Bad Request**: Invalid request parameters.
- **401 Unauthorized**: Authentication failed.
- **403 Forbidden**: Insufficient permissions.
- **404 Not Found**: Resource not found.
- **500 Internal Server Error**: Server error.

### Notes
- Request frequency limit: QPS per account must not exceed 60.
- The `200 OK` response only indicates that the agent task has been successfully dispatched, not that the agent has joined the room or is operational. Monitor agent status via event callbacks.
- If human users leave the room, the agent task will automatically stop after 180 seconds (billable during this period). Use `AgentConfig.IdleTimeout` to customize this duration. Call `StopVoiceChat` to terminate tasks promptly and avoid unnecessary costs.
```

--------------------------------

### Third-Party LLM/Agent Configuration - Tool Definitions

Source: https://www.volcengine.com/docs/6348/2123348/index

Specifies tools, specifically functions, that the model can call. This includes the tool type, function name, description, and parameters conforming to JSON Schema.

```json
{
  "Tools": [
    {
      "type": "function",
      "function": {
        "name": "get_current_weather",
        "description": "Get the weather information for a specified city.",
        "parameters": {
          "type": "object",
          "properties": {
            "location": {
              "type": "string",
              "description": "The city to query weather for, e.g., 'San Francisco'"
            },
            "date": {
              "type": "string",
              "description": "The date for the weather query in YYYY-MM-DD format. Defaults to today."
            }
          },
          "required": ["location"]
        }
      }
    }
  ]
}
```

--------------------------------

### Volcengine Web Search Agent Configuration

Source: https://www.volcengine.com/docs/6348/2123348/index

Configure the web search agent for Volcengine, enabling it and providing necessary parameters like API Key, ParamsString, FunctionName, FunctionDescription, and ComfortWords.

```json
{
  "Enable": true,
  "APIKey": "your_agent_apikey",
  "ParamsString": "{\"bot_id\": \"742...\", \"stream\": true}",
  "FunctionName": "WebSearch",
  "FunctionDescription": "查询实时信息，如今天的天气、最新的新闻、A 股票的当前价格等",
  "ComfortWords": "正在为您联网查询，请稍等..."
}
```

--------------------------------

### StartVoiceChat API

Source: https://www.volcengine.com/docs/6348/2123348/index

Initiates a voice chat session. This endpoint allows for detailed configuration of Automatic Speech Recognition (ASR), Large Language Model (LLM), and Text-to-Speech (TTS) components, along with agent-specific settings.

```APIDOC
## POST /StartVoiceChat

### Description
Initiates a voice chat session with configurable ASR, LLM, and TTS settings, along with agent-specific behaviors like burst audio and voiceprint recognition.

### Method
POST

### Endpoint
`https://rtc.volcengineapi.com?Action=StartVoiceChat&Version=2025-06-01`

### Parameters
#### Path Parameters
None

#### Query Parameters
- **Action** (string) - Required - The action to perform, must be 'StartVoiceChat'.
- **Version** (string) - Required - The API version, must be '2025-06-01'.

#### Request Body
- **AppId** (string) - Required - The application ID.
- **RoomId** (string) - Required - The ID of the chat room.
- **TaskId** (string) - Required - The ID of the chat task.
- **Config** (object) - Optional - Configuration for ASR, LLM, TTS, and other features.
  - **ASRConfig** (object) - Optional - Automatic Speech Recognition configuration.
    - **Provider** (string) - Optional - ASR provider (e.g., 'volcano').
    - **ProviderParams** (object) - Optional - Parameters specific to the ASR provider.
      - **AccessToken** (string) - Optional - Access token for the ASR provider. Not required when using large model streaming ASR.
      - **Mode** (string) - Optional - ASR mode (e.g., 'bigmodel').
      - **StreamMode** (integer) - Optional - Stream mode for ASR.
    - **VADConfig** (object) - Optional - Voice Activity Detection configuration.
      - **SilenceTime** (integer) - Optional - Duration of silence to detect end of speech.
      - **AIVAD** (boolean) - Optional - Enable AI-based VAD.
    - **InterruptConfig** (object) - Optional - Configuration for interrupting speech.
  - **LLMConfig** (object) - Optional - Large Language Model configuration.
    - **Mode** (string) - Optional - LLM mode (e.g., 'ArkV3').
    - **ModelName** (string) - Optional - Name of the LLM model.
    - **SystemMessages** (array of strings) - Optional - System messages to guide the LLM's behavior.
    - **HistoryLength** (integer) - Optional - Number of historical turns to consider.
    - **ThinkingType** (string) - Optional - Type of thinking process for the LLM.
    - **VisionConfig** (object) - Optional - Vision-related configuration.
      - **SnapshotConfig** (object) - Optional - Configuration for taking snapshots.
      - **StorageConfig** (object) - Optional - Configuration for storing vision data.
        - **TosConfig** (object) - Optional - TOS storage configuration.
  - **TTSConfig** (object) - Optional - Text-to-Speech configuration.
    - **Provider** (string) - Optional - TTS provider (e.g., 'volcano_bidirection').
    - **ProviderParams** (object) - Optional - Parameters specific to the TTS provider.
      - **AccessToken** (string) - Optional - Access token for the TTS provider. Not required when using large model voice synthesis.
      - **Credential** (object) - Optional - TTS credential information.
        - **ResourceId** (string) - Optional - Resource ID for TTS.
      - **VolcanoTTSParameters** (string) - Optional - JSON string of Volcano TTS specific parameters.
  - **SubtitleConfig** (object) - Optional - Subtitle configuration.
  - **InterruptMode** (integer) - Optional - Mode for interrupting TTS.
  - **FunctionCallingConfig** (object) - Optional - Configuration for function calling.
  - **WebSearchAgentConfig** (object) - Optional - Configuration for web search agent.
  - **MemoryConfig** (object) - Optional - Configuration for memory.
  - **MusicAgentConfig** (object) - Optional - Configuration for music agent.
- **AgentConfig** (object) - Optional - Agent-specific configuration.
  - **TargetUserId** (array of strings) - Optional - User IDs targeted by the agent.
  - **UserId** (string) - Optional - The ID of the agent.
  - **WelcomeMessage** (string) - Optional - The initial welcome message from the agent.
  - **Burst** (object) - Optional - Configuration for burst audio sending.
    - **Enable** (boolean) - Optional - Enable or disable burst audio. Defaults to false.
    - **BufferSize** (integer) - Optional - Size of the initial audio buffer in milliseconds. Defaults to 500.
    - **Interval** (integer) - Optional - Interval for sending subsequent audio data in milliseconds. Defaults to 10.
  - **IdleTimeout** (integer) - Optional - Timeout in seconds for idle waiting before the agent exits. Defaults to 180.
  - **AnsMode** (integer) - Optional - AI noise reduction mode. 0: Disabled, 1: Light, 2: Medium, 3: Heavy. Defaults to 0.
  - **VoicePrint** (object) - Optional - Voiceprint configuration.
    - **MetaList** (array) - Optional - Deprecated. Use VoicePrintList.
    - **VoicePrintList** (object) - Optional - List of voiceprint IDs and configurations.
      - **Mode** (integer) - Optional - Voiceprint processing mode. 0: Disabled, 1: Noise Reduction, 2: Recognition. Defaults to 0.
      - **IdList** (array of strings) - Optional - List of pre-registered voiceprint IDs. Required for Mode 1 (with pre-registered voiceprints) and Mode 2.
      - **Score** (integer) - Optional - Confidence threshold for voiceprint recognition (1-100). Only effective when Mode is 2 and IdList is provided. Defaults to 50.
      - **EnableSV** (boolean) - Optional - Enable speaker verification when using pre-registered voiceprints for noise reduction. Only effective when Mode is 1 and IdList is provided. Defaults to false.

### Request Example
```json
{
  "AppId": "68e*****477",
  "RoomId": "ChatRoom01",
  "TaskId": "ChatTask01", 
  "Config": {
    "ASRConfig": {
      "Provider": "volcano",
      "ProviderParams": {
        "Mode": "bigmodel",
        "StreamMode": 0
      },
      "VADConfig": {
        "SilenceTime": 600,
        "AIVAD": true
      },
      "InterruptConfig": {}
    },
    "LLMConfig": {
      "Mode": "ArkV3",
      "ModelName": "doubao-seed-1-8-251215",
      "SystemMessages": [
        "##人设\n你是一个全能的超级助手，具备强大的知识库、情感理解能力和解决问题的能力。你的目标是高效、专业、友好地帮助用户完成各类任务，包括但不限于日常生活、工作安排、信息检索、学习辅导、创意写作、语言翻译和技术支持等；\n\n##约束\n始终主动、礼貌、有条理；\n回答准确但不冗长，必要时可提供简洁总结+详细解释；\n不清楚的任务会主动澄清，不假设、不误导。"
      ],
      "HistoryLength": 10,
      "ThinkingType": "disabled",
      "VisionConfig": {
        "SnapshotConfig": {},
        "StorageConfig": {
          "TosConfig": {}
        }
      }
    },
    "TTSConfig": {
      "Provider": "volcano_bidirection",
      "ProviderParams": {
        "Credential": {
          "ResourceId": "seed-tts-2.0"
        },
        "VolcanoTTSParameters": "{\"req_params\":{\"speaker\":\"ICL_zh_female_wuxi_tob\",\"audio_params\":{\"speech_rate\":0}}}"
      }
    },
    "SubtitleConfig": {},
    "InterruptMode": 0,
    "FunctionCallingConfig": {},
    "WebSearchAgentConfig": {},
    "MemoryConfig": {},
    "MusicAgentConfig": {}
  },
  "AgentConfig": {
    "TargetUserId": ["Huoshan01"],
    "UserId": "ChatBot01",
    "WelcomeMessage": "我是你的AI助手，有什么需要我为您效劳的吗？",
    "Burst": {
      "Enable": false,
      "BufferSize": 0,
      "Interval": 0
    },
    "VoicePrint": {
      "MetaList": null,
      "VoicePrintList": null
    }
  }
}
```

### Response
#### Success Response (200)
- **Result** (string) - 'ok' if the request is successful, empty otherwise.
- **ResponseMetadata** (object) - Metadata about the response.
  - **RequestId** (string) - The ID of the request.
  - **Action** (string) - The action performed.
  - **Version** (string) - The API version used.
  - **Service** (string) - The service name.
  - **Region** (string) - The region where the request was processed.

#### Response Example
```json
{
    "Result": "ok",
    "ResponseMetadata": {
        "RequestId": "202512****10420",
        "Action": "StartVoiceChat",
        "Version": "2025-06-01",
        "Service": "rtc",
        "Region": "cn-north-1"
    }
}
```
```

--------------------------------

### Configure TTS Parameters (Volcano Bidirectional - JSON String)

Source: https://www.volcengine.com/docs/6348/2123348/index

Passes native API parameters for Volcengine's bidirectional streaming TTS service as a JSON string. This includes mandatory speaker selection and optional parameters like speech rate. Certain platform-managed parameters are not allowed.

```json
{"req_params":{"speaker":"zh_female_linjianvhai_moon_bigtts","audio_params":{"speech_rate":100}}}
```

--------------------------------

### Web Search Agent Configuration

Source: https://www.volcengine.com/docs/6348/2123348/index

Configure settings for the web search agent.

```APIDOC
## Web Search Agent Configuration

### Description
Configuration for the web search agent, enabling the AI to perform web searches.

### Parameters
#### Request Body
- **Enable** (Boolean) - Optional - Whether to enable the web search capability. Defaults to `false`.
- **APIKey** (String) - Optional - The API Key for the web search agent service. Required if `Enable` is `true`.
- **ParamsString** (String) - Optional - Transmitted parameters for the web search agent service (compressed and escaped JSON string). Required if `Enable` is `true`. This should be a JSON string conforming to the ChatCompletionRequest parameters, including `bot_id` and `stream: true`.
- **FunctionName** (String) - Optional - The name of the web search tool. Required if `Enable` is `true`. Must not conflict with other tool names.
- **FunctionDescription** (String) - Optional - A description of the web search tool for the AI. Required if `Enable` is `true`.
- **ComfortWords** (String) - Optional - A phrase to be spoken by the AI while performing a web search. If empty, no comfort words are spoken.
```

--------------------------------

### Configure MiniMax TTS Settings (JSON)

Source: https://www.volcengine.com/docs/6348/2123348/index

This snippet shows the configuration for the MiniMax TTS service. It requires authentication credentials, including an API key and group ID, and allows selection of the model version for speech synthesis.

```json
{
  "Authorization": "eyJhbG****SUzI1N",
  "Groupid": "983*****669",
  "model": "speech-01-turbo"
}
```

--------------------------------

### Agent Configuration

Source: https://www.volcengine.com/docs/6348/2123348/index

Core configuration for the AI agent, including user IDs, welcome messages, and callback settings.

```APIDOC
## Agent Configuration

### Description
Core settings for the AI agent, including target user, welcome message, agent ID, and conversation state callbacks.

### Parameters
#### Request Body
- **TargetUserId** (String[]) - Required - The User ID(s) of the human user(s) interacting with the agent. Only one User ID is supported for one-on-one calls.
- **WelcomeMessage** (String) - Optional - The welcome message spoken by the agent when it starts.
- **UserId** (String) - Optional - The ID of the agent. Defaults to `voiceChat_$(TargetUserId)_$(timestamp_now)` if not provided. Should be unique within the same AppId.
- **EnableConversationStateCallback** (Boolean) - Optional - Whether to receive agent state change messages (e.g., 'listening', 'thinking', 'speaking'). Defaults to `false`.
- **ServerMessageURLForRTS** (String) - Optional - The URL of your business server to receive agent state change messages. Required if `EnableConversationStateCallback` is `true` and server-side callbacks are desired. Must be publicly accessible and capable of handling POST requests.
- **ServerMessageSignatureForRTS** (String) - Optional - A custom signature key to verify the authenticity of callback requests. Required if `EnableConversationStateCallback` is `true` and server-side callbacks are desired.
- **UseLicense** (Boolean) - Optional - Whether this is a License user. Defaults to `false`. If `true`, requires enabling a whitelist and obtaining ASR, TTS, and LLM parameters from the console.
```

--------------------------------

### Configure Voice Cloning (Parameter Passthrough) (JSON)

Source: https://www.volcengine.com/docs/6348/2123348/index

This configuration enables advanced control for Volcano Engine's voice cloning service using parameter passthrough. It allows flexible customization by directly passing native API parameters, including specifying the speaker ID and audio parameters like speech rate.

```json
{
  "VolcanoTTSParameters": "{\"req_params\":{\"speaker\":\"S_The****K1\",\"audio_params\":{\"speech_rate\":100}}}",
  "Credential": {
    "ResourceId": "seed-icl-2.0"
  }
}
```

--------------------------------

### MCP Configuration

Source: https://www.volcengine.com/docs/6348/2123348/index

Configure MCP settings for integrating with external tools and services.

```APIDOC
## POST /websites/volcengine_6348_2123348

### Description
Configure MCP settings for accessing external tool services like knowledge bases. Ensure your MCP Server supports the streaming SSE protocol and configure custom headers in `LLMConfig.ExtraHeader` if necessary.

### Method
POST

### Endpoint
/websites/volcengine_6348_2123348

### Parameters
#### Request Body
- **URL** (String) - Required - The access address of the remote MCP Server.
- **Name** (String) - Required - A unique name to identify the MCP service. This name is used by the LLM to distinguish tool sources and route requests to the corresponding MCP Server. It must not conflict with names defined in `WebSearchAgentConfig.FunctionName` or `Tools.function.name`.
- **MCP** (Object[]) - Optional - Configuration for MCP, used for accessing external tool services.
  - **URL** (String) - Required - The access address of the remote MCP Server.
  - **Name** (String) - Required - A unique name to identify the MCP service.
  - **ComfortWords** (String) - Optional - Comfort word content played to the user when the agent calls an MCP tool. Defaults to empty.
  - **InterestedTools** (String[]) - Optional - A list of tools to be added to the LLM context for decision-making and invocation. If empty or not provided, all tools offered by the MCP Server are added. If specified, only the listed tools are added.
  - **Feature** (String) - Optional - Used for testing with HTTP domains. Fixed value: `{"Http":true}`.
  - **Prefill** (Boolean) - Optional - Whether to pre-fill intermediate ASR results into the LLM for processing to reduce latency. Defaults to `false`.
  - **VisionConfig** (Object) - Optional - Configuration for visual understanding capabilities, supporting RTC live video streams or external images.
    - **Enable** (Boolean) - Optional - Whether to enable visual understanding. Defaults to `false`.
    - **SnapshotConfig** (Object) - Optional - Configuration for frame grabbing and screenshotting. The agent automatically grabs frames based on the configured strategy.
      - **StreamType** (Integer) - Optional - Screenshot stream type: 0 for main stream (camera), 1 for screen sharing. Defaults to 0.
      - **ImageDetail** (String) - Optional - Image detail setting. Defaults to 'auto'.
      - **Height** (Integer) - Optional - Video frame height sent to the LLM. Range: [0, 1792] pixels. Defaults to 360 if 0 or not provided. Width is calculated based on height.
      - **Interval** (Integer) - Optional - Time interval for screenshots. Defaults to '' (empty).
      - **ImagesLimit** (Integer) - Optional - Number of images sent to the LLM per request. Range: [0, 50]. Defaults to 2 if 0 or not provided.
  - **StorageConfig** (Object) - Optional - Storage configuration.
    - **Type** (Integer) - Optional - Storage type: 0 for Base64 encoding (local, auto-deleted after session), 1 for TOS. Defaults to 0.
    - **TosConfig** (Object) - Optional - TOS configuration.
      - **AccountId** (String) - Optional - User account ID.
      - **Region** (Integer) - Optional - Region for TOS storage. Defaults to 0.
      - **Bucket** (String) - Optional - Bucket name for TOS storage.
  - **EnableRoundId** (Boolean) - Optional - Whether to include `round_id` (conversation round ID) in requests to third-party models/Agents. Defaults to `false`.
  - **Custom** (String) - Optional - Custom JSON string for passing business-defined parameters. Defaults to empty.
  - **ExtraHeader** (JSONMap) - Optional - Custom passthrough headers. A JSON object whose key-value pairs will be forwarded as additional HTTP header fields to your custom model service request.
  - **SubtitleConfig** (Object) - Optional - Configuration for obtaining real-time subtitles (dialogue records).
    - **DisableRTSSubtitle** (Boolean) - Optional - Whether to disable client-side subtitle callbacks. Defaults to `false`.
    - **ServerMessageUrl** (String) - Optional - Your business server address (URL) for receiving subtitle messages. Required if receiving subtitles server-side.
    - **ServerMessageSignature** (String) - Optional - Signature key to verify the authenticity of callback requests. Required if receiving subtitles server-side.
    - **SubtitleMode** (Integer) - Optional - Whether to align subtitle callbacks with audio timestamps. 0 for alignment, 1 for no alignment (faster callback). Defaults to 0.
  - **FunctionCallingConfig** (Object) - Optional - Configuration for receiving Function Calling invocation notifications and instructions via your business server.
    - **ServerMessageUrl** (String) - Required - Your business server address (URL) for receiving Function Calling notifications and instructions.
    - **ServerMessageSignature** (String) - Required - Signature key to verify the authenticity of callback requests.
  - **InterruptMode** (Integer) - Optional - Whether to enable voice interruption (interrupt upon speaking): 0 for enable, 1 for disable. Defaults to 0.

### Request Example
```json
{
  "URL": "https://knowledge-mcp-server.com/test/123",
  "Name": "knowledge",
  "ComfortWords": "正在处理中...",
  "InterestedTools": ["search_knowledge"],
  "Feature": "{\"Http\":true}",
  "Prefill": false,
  "VisionConfig": {
    "Enable": true,
    "SnapshotConfig": {
      "StreamType": 0,
      "ImageDetail": "auto",
      "Height": 640,
      "Interval": 1000,
      "ImagesLimit": 2
    }
  },
  "StorageConfig": {
    "Type": 0,
    "TosConfig": {
      "AccountId": "account_id",
      "Region": 0,
      "Bucket": "bucket"
    }
  },
  "EnableRoundId": false,
  "Custom": "{}",
  "ExtraHeader": {"custom_header": "value"},
  "SubtitleConfig": {
    "DisableRTSSubtitle": false,
    "ServerMessageUrl": "https://example-domain.com/vertc/subtitle",
    "ServerMessageSignature": "b46ab5f1d8ad6a",
    "SubtitleMode": 0
  },
  "FunctionCallingConfig": {
    "ServerMessageUrl": "https://example-domain8080/m2",
    "ServerMessageSignature": "TestSignature"
  },
  "InterruptMode": 0
}
```

### Response
#### Success Response (200)
- **message** (String) - Success message.

#### Response Example
```json
{
  "message": "MCP configuration updated successfully."
}
```
```

--------------------------------

### Volcengine Digital Human Configuration

Source: https://www.volcengine.com/docs/6348/2123348/index

Configure the Volcengine digital human feature, including enabling it, setting the AppID, Token, Type, Role, background URL, and video bitrate.

```json
{
  "Enabled": true,
  "AvatarAppID": "zt****ubf",
  "AvatarToken": "zOpP1FFZ****ZQi9GHn",
  "AvatarType": "3min",
  "AvatarRole": "250623-****-linyunzhi",
  "BackgroundUrl": "https://tos-tools.tos-cn-beijing.volces.com/misc/sample.png",
  "VideoBitrate": 3000
}
```
Tokens:
Raw