StreamingTTSClientConfig
========================

**class** ``StreamingTTSClientConfig`` **(** :doc:`BaseClientConfig` **)**

    Provider-agnostic paced text-in/audio-out WebSocket TTS configuration.

**Polymorphic Type:**

    ``type: streaming_tts``

    All ``BaseClientConfig`` types:

    - ``openai_chat_completions``: :doc:`OpenAIChatCompletionsClientConfig`
    - ``tts``: :doc:`TTSClientConfig`
    - ``streaming_tts``: :doc:`StreamingTTSClientConfig`
    - ``stt``: :doc:`STTClientConfig`
    - ``openai_completions``: :doc:`OpenAICompletionsClientConfig`
    - ``openai_router``: :doc:`OpenAIRouterClientConfig`

**Fields:**

``api_base`` : *Optional* [ *str* ] = ``None``
    API base URL. Defaults to OPENAI_API_BASE env var.

``api_key`` : *Optional* [ *str* ] = ``None``
    API key. Defaults to OPENAI_API_KEY env var.

``model`` : *str* = ``""``
    The streaming TTS model ID.

``address_append_value`` : *str* = ``"chat/completions"``
    The address append value for the LLM API.

``request_timeout`` : *int* = ``300``
    The timeout for each request to the LLM API (in seconds).

``additional_sampling_params`` : *str* = ``"{}"``
    Additional sampling params to send with each request to the LLM API.

``provider`` : *str* = ``""``
    Streaming TTS provider. Supported: openai_realtime, vajra, elevenlabs, deepgram_flux, deepgram_aura, cartesia.

``voice_id`` : *str* = ``""``
    Optional provider voice identifier.

``sample_rate`` : *int* = ``24000``
    PCM sample rate in Hz.

``api_key_env`` : *Optional* [ *str* ] = ``None``
    Optional provider API-key environment variable override.

``pacing`` : :doc:`TextPacingConfig` = :doc:`TextPacingConfig`
    Upstream LLM text pacing configuration.

``abort`` : :doc:`TTSAbortConfig` = :doc:`TTSAbortConfig`
    Adversarial mid-stream abort injection for the Vajra provider.

``input_output_mode`` : *str* = ``"complete_text"``
    Explicit response scheduling: complete_text or duplex. Providers that synthesize on text receipt use their native trigger behavior.

``duplex_start_after_tokens`` : *int* = ``1``
    Whitespace-delimited input words before an explicit duplex response trigger. The field name is retained for compatibility.

``language`` : *Optional* [ *str* ] = ``None``
    Optional language for protocols that support it.

``instructions`` : *Optional* [ *str* ] = ``None``
    Optional synthesis instructions for protocols that support them.

``task_type`` : *Optional* [ *str* ] = ``None``
    Optional provider task type.

``chunk_length_schedule`` : *list* [ *int* ] = ``[120, 160, 250, 290]``
    ElevenLabs character thresholds for audio generation.

``stability`` : *float* = ``0.5``
    ElevenLabs voice stability.

``similarity_boost`` : *float* = ``0.8``
    ElevenLabs similarity boost.

``speed`` : *float* = ``1.0``
    Provider speaking-rate multiplier.

``auto_mode`` : *bool* = ``False``
    Enable ElevenLabs automatic chunk scheduling.

``apply_text_normalization`` : *str* = ``"off"``
    ElevenLabs text normalization mode: auto | on | off.

``mip_opt_out`` : *bool* = ``False``
    Opt out of the Deepgram Model Improvement Program.

``cartesia_version`` : *str* = ``"2026-03-01"``
    Cartesia API version header.

``max_buffer_delay_ms`` : *Optional* [ *int* ] = ``None``
    Optional Cartesia server-side transcript buffer cap in milliseconds. Set it explicitly for reproducible benchmark runs.