Interface InworldSTTOptions

interface InworldSTTOptions {
    audio_encoding?: "LINEAR16" | "AUTO_DETECT";
    enable_voice_profile?: boolean;
    end_of_turn_confidence_threshold?: number;
    inactivity_timeout_seconds?: number;
    include_word_timestamps?: boolean;
    min_end_of_turn_silence_when_confident?: number;
    prompts?: string[];
    vad_threshold?: number;
    voice_profile_top_n?: number;
}

Properties

audio_encoding? enable_voice_profile? end_of_turn_confidence_threshold? inactivity_timeout_seconds? include_word_timestamps? min_end_of_turn_silence_when_confident? prompts? vad_threshold? voice_profile_top_n?

Properties

Optional audio_encoding

audio_encoding?: "LINEAR16" | "AUTO_DETECT"

Wire-format encoding sent to Inworld. Default: LINEAR16.

Optional enable_voice_profile

enable_voice_profile?: boolean

Enable Voice Profile detection. Default: true.

Optional end_of_turn_confidence_threshold

end_of_turn_confidence_threshold?: number

End-of-turn confidence threshold (0.0–1.0). Default: 0.5.

Optional inactivity_timeout_seconds

inactivity_timeout_seconds?: number

Stop transcription after this many seconds of silence; 0 disables.

Optional include_word_timestamps

include_word_timestamps?: boolean

Enable word-level timestamps. Default: true.

Optional min_end_of_turn_silence_when_confident

min_end_of_turn_silence_when_confident?: number

Minimum end-of-turn silence in milliseconds when confident.

Optional prompts

prompts?: string[]

Domain-specific contextual hints passed to the model.

Optional vad_threshold

vad_threshold?: number

VAD threshold (0.0–1.0). Default: 0.5.

Optional voice_profile_top_n

voice_profile_top_n?: number

Max labels per category in voice-profile responses (1–20). Default: 10.

Settings

Member Visibility

Theme

On This Page

audio_encodingenable_voice_profileend_of_turn_confidence_thresholdinactivity_timeout_secondsinclude_word_timestampsmin_end_of_turn_silence_when_confidentpromptsvad_thresholdvoice_profile_top_n