AIAgentConfig

The ID of the call ambient sound. You can obtain the ID from the advanced settings of the agent configuration in the console.

The volume of the call background sound. Valid values: 0 to 100. A value of 0 indicates that the background sound is disabled.

The ASR context enhancement.

The ASR language ID. Valid values:

• zh_mandarin: Chinese.

• en: English.

• zh_en: Mixed Chinese and English.

• es: Spanish.

• jp: Japanese.

The threshold for speech segmentation detection. If the silence duration exceeds this threshold, the speech is considered segmented. Valid values: 200 to 1200. Unit: milliseconds. Default value: 400.

The instance-level real-time hotwords and their weights for ASR.

The pass-through parameters for custom ASR integration.

The minimum duration threshold for voice activity detection (VAD), which is used to control the sensitivity of interruptions. A value of 0 indicates that this feature is disabled. Valid values: 200 to 2000. Unit: milliseconds. Common values range from 200 to 500, which correspond to 1 to 4 words. Default value: null. If this parameter is left empty, this feature does not take effect.

The interruption threshold parameter. Valid values: 0 to 11. Default value: 11.

• 0: Disables the VAD feature.

• 1 to 10: A larger value indicates that it is more difficult to trigger an interruption.

• 11: Significantly different from previous values. It causes less damage to the pre-processed conversation audio and provides stronger anti-interference capability.

The model ID of the digital human.

The resolution of the digital human video stream. The value must be in strictly lowercase letters and must be specified together with AvatarId. Valid values:

• 360p

• 480p

• 720p

• 1080p

The ID of the digital human template, which specifies the template used when the digital human is started. The template is provided by the digital human service provider, and the provider also verifies whether the specified value is valid.

Specifies whether to support speech interruption. Default value: true.

Specifies whether to retain interrupt words when sending messages to the LLM. By default, interrupt words are discarded. For example, if the interrupt word is "wait", the message "wait how is the weather today" is changed to "how is the weather today" after the interrupt word is discarded.

The ASR processing policy in No Interrupting mode. Valid values:

• cache: Caches the ASR text. After the current turn ends, the cached ASR text is processed together in the next turn.

• discard: Directly discards the ASR text. Default value: cache.

The parameters for Alibaba Cloud Model Studio. The value must be a JSON string. For more information about the parameter format, see Alibaba Cloud Model Studio parameters.

Specifies whether to keep the LLM message history consistent with the text-to-speech (TTS) playback content. Default value: false. After this feature is enabled, the saved LLM messages are kept consistent with the TTS playback content. When the user interrupts the agent in the LLM message history, the <ims_agent_interrupted> tag is inserted at the interruption position in the next message sent to the LLM. Example:

[
  {"role": "user", "content": "Tell me a story."},
  {"role": "assistant", "content": "Okay, let me tell you a story about Romance of the Three Kingdoms. Do<ims_agent_interrupted> you want to listen?"},
  {"role": "user", "content": "Tell me another one."}
]

Specifies whether to send the complete LLM result to the client after the LLM generates the complete response. This switch does not affect the streaming generation of subtitles.

The maximum number of retained turns for the historical conversation context of the LLM or multimodal LLM. Default value: 10.

The system prompt for the LLM after the call is started.

The extra query parameters for the OpenAI-compatible LLM. Parameters must be in the key=value format. Multiple parameters are connected by an ampersand (&). All values must be of the string type.

The maximum delay time for text output in milliseconds. If this time is exceeded, the cached text is forcibly output. Valid values: 1000 to 10000. A value of 0 or empty indicates no limit. Default value: empty.

The minimum length of text output in characters. Text shorter than this length is cached and concatenated. Valid values: 0 to 100. A value of 0 or empty indicates no limit. Default value: empty.

The default transfer number. This is a backup number. You must still explicitly pass the called number when you initiate a transfer call.

The caller number used for transfer calls. Configuring this field enables the RTC transfer capability. If this field is not configured, the behavior remains unchanged.

Specifies whether to perform real-time speech recognition for both parties (the RTC user and the transferred phone party) after the transfer is connected. When enabled, the recognition results are pushed to the client as subtitles through the RTC message channel and written to the call record. Valid values:

• true: Enabled. This is the default value.

• false: Disabled. Subtitles and call records are not generated during the transfer.

The custom SIP outbound gateway address. This parameter is required only when Type is set to 1.

Valid values:

• 0: Alibaba Cloud communication line. This is the default value.

• 1: Custom SIP line.

The emotion of the voice. Currently, only MiniMax is supported. MiniMax supports the following seven emotions:

• happy: Happy

• sad: Sad

• angry: Angry

• fearful: Fearful

• disgusted: Disgusted

• surprised: Surprised

• calm: Calm

The TTS pronunciation correction and text replacement configuration.

The natural language instruction control for TTS.

The language ID. Currently, only MiniMax is supported. Default value: empty. This parameter enhances the recognition capability for specified minority languages and dialects. Setting this parameter improves the voice performance in the specified minority language or dialect scenarios. If the minority language type is not specified, you can set this parameter to auto, and the model automatically determines the language type. Valid values:

The model ID. Currently, only MiniMax is supported. Valid values: speech-01-turbo and speech-02-turbo.

The speech rate. This parameter is supported on all platforms. For CosyVoice, the default value is 1.0, and the valid values range from 0.5 to 2.0. For MiniMax, the default value is 1.0, and the valid values range from 0.5 to 2.0.

The voice ID. Modifications take effect in the next sentence. If you do not specify this parameter, the voice ID configured in the agent template is used. This parameter takes effect only for preset TTS. The input length cannot exceed 64 characters. For valid values, refer to Intelligent speech effect examples.

The eagerness level. This parameter takes effect only in Semantic mode. It controls how quickly the AI starts to respond after detecting a pause:

• Low: Patient waiting. The maximum wait time is 6 seconds, which reduces the risk of being interrupted.

• Medium: Balanced waiting. The maximum wait time is 4 seconds, which is suitable for most scenarios.

• High: Fast response. The maximum wait time is 2 seconds, which improves speed but may increase the risk of false cut-offs. Default value: empty.

The turn detection mode. Valid values:

• Normal: Normal mode. AI is not used to determine semantics. This is the default value.

• Semantic: Semantic mode. AI is used to determine whether the user has finished speaking based on the context semantics.

The pause determination time in AI mode. Unit: milliseconds. Default value: -1. Valid values:

• -1: The AI automatically determines the appropriate wait time.

• 0 to 10000: A custom wait time. We recommend that you set this value to 0 to 1500 milliseconds.

The voiceprint registration mode. Default value: Explicit.

Specifies whether to use voiceprint recognition. Default value: false. When voiceprint recognition is enabled, a valid voiceprint ID must be passed.

The unique identity ID for voiceprint recognition. Default value: empty. The passed voiceprint ID must be registered through the voiceprint registration API operation. For more information about the API operation, refer to Register human voiceprint.