AI Assistant Voice Configuration

Configure voice-specific settings for assistants with assistant_type set to voice.

Endpoint

PATCH

/ai-assistants/:id

Request

Response

json

Voice Configuration Fields

Field	Type	Description
`acting_user_id`	Number or null	User used for voice assistant write actions. If omitted or null, the creator/default user is used
`telnyx_ai_assistant_config.sync_status`	String	Set to `pending` after changing voice settings so the configuration can sync
`telnyx_ai_assistant_config.voice_runtime`	String	Runtime used by the voice assistant. UI values include `telnyx_builtin`, `custom_llm`, and `openai_realtime_sip`
`telnyx_ai_assistant_config.use_custom_llm`	Boolean	Whether the assistant uses a custom LLM runtime
`telnyx_ai_assistant_config.model`	String	Telnyx built-in model ID, such as `openai/gpt-4.1`
`telnyx_ai_assistant_config.openai_realtime_model`	String	OpenAI realtime model when using the realtime SIP runtime
`telnyx_ai_assistant_config.greeting`	String	Initial spoken greeting
`telnyx_ai_assistant_config.voice_provider`	String	TTS provider, such as `telnyx`, `elevenlabs`, `aws`, `azure`, `rime`, `resemble`, `minimax`, or `inworld`
`telnyx_ai_assistant_config.voice_provider_model`	String or null	Provider-specific voice model
`telnyx_ai_assistant_config.voice_id`	String	Provider voice ID
`telnyx_ai_assistant_config.voice_name`	String	Display name for the selected voice
`telnyx_ai_assistant_config.voice_model`	String	Full selected voice identifier
`telnyx_ai_assistant_config.voice_api_key`	String	Provider API key when required by the selected TTS provider
`telnyx_ai_assistant_config.voice_api_key_ref`	String	Stored provider key reference returned by the system
`telnyx_ai_assistant_config.voice_speed`	Number	Speech speed multiplier
`telnyx_ai_assistant_config.voice_language_boost`	String	Language boost, such as `English` or `Spanish`
`telnyx_ai_assistant_config.voice_expressive_mode`	Boolean	Whether expressive mode is enabled for supported voices
`telnyx_ai_assistant_config.voice_similarity_boost`	Number	ElevenLabs similarity boost
`telnyx_ai_assistant_config.voice_style`	Number	ElevenLabs style value
`telnyx_ai_assistant_config.voice_use_speaker_boost`	Boolean	ElevenLabs speaker boost
`telnyx_ai_assistant_config.background_audio`	Object or null	Background audio settings
`telnyx_ai_assistant_config.transcription_model`	String	Speech-to-text model
`telnyx_ai_assistant_config.transcription_language`	String	Language code or `auto`
`telnyx_ai_assistant_config.speaking_wait_seconds`	Number	Minimum wait before speaking
`telnyx_ai_assistant_config.speaking_on_punctuation_seconds`	Number	Wait after punctuation
`telnyx_ai_assistant_config.speaking_on_no_punctuation_seconds`	Number	Wait when no punctuation is detected
`telnyx_ai_assistant_config.speaking_on_number_seconds`	Number	Wait around numbers
`telnyx_ai_assistant_config.interruption_enable`	Boolean	Whether callers can interrupt the assistant
`telnyx_ai_assistant_config.transfer_from_number`	String or null	Company number used for external transfers when company caller ID is selected
`telnyx_ai_assistant_config.external_transfer_caller_id_mode`	String	`original_caller` or `company_number`
`telnyx_ai_assistant_config.widget_settings`	Object or null	Embedded web widget appearance settings

Background Audio

json

Accepted type values used by the configuration UI are:

Type	Description
`predefined_media`	Use a predefined background audio option, such as `silence` or `office`
`media_url`	Use a public audio URL

Web Widget Settings

json

When the voice assistant has a synced Telnyx assistant ID, embed it on your website with:

html

External Voice Handoff

Voice assistants can transfer calls to external numbers or SIP URIs through external_handoff_targets on the assistant and caller ID fields inside telnyx_ai_assistant_config.

List Telnyx Built-in Models

GET

/telnyx-ai-models

List Transcription Languages

GET

/telnyx-ai-transcription-languages

List or Load Voices

GET

/telnyx-ai-voices

Some providers can require an API key before voices are loaded. Use

POST

/telnyx-ai-voices when you need to load provider voices with provider-specific credentials.

Sync From Telnyx

GET

/ai-assistants/:id

Use this endpoint after external Telnyx changes when you want to refresh the stored voice configuration.

Report an issue with this documentation

Please log in to report issues with our documentation.