Struct RealtimeConfig 

pub struct RealtimeConfig {

    pub model: Option<String>,
    pub instruction: Option<String>,
    pub voice: Option<String>,
    pub modalities: Option<Vec<String>>,
    pub input_audio_format: Option<AudioEncoding>,
    pub output_audio_format: Option<AudioEncoding>,
    pub turn_detection: Option<VadConfig>,
    pub tools: Option<Vec<ToolDefinition>>,
    pub tool_choice: Option<String>,
    pub input_audio_transcription: Option<TranscriptionConfig>,
    pub temperature: Option<f32>,
    pub max_response_output_tokens: Option<u32>,
    pub cached_content: Option<String>,
    pub interruption_detection: Option<InterruptionDetection>,
    pub affective_dialog: Option<bool>,
    pub extra: Option<Value>,
}

Available on crate feature realtime only.

Configuration for a realtime session.

Fields

model: Option<String>

Model to use (provider-specific).

instruction: Option<String>

System instruction for the agent.

voice: Option<String>

Voice to use for audio output.

modalities: Option<Vec<String>>

Output modalities: [“text”], [“audio”], or [“text”, “audio”].

input_audio_format: Option<AudioEncoding>

Input audio format.

output_audio_format: Option<AudioEncoding>

Output audio format.

turn_detection: Option<VadConfig>

VAD configuration.

tools: Option<Vec<ToolDefinition>>

Available tools/functions.

tool_choice: Option<String>

Tool selection mode: “auto”, “none”, “required”.

input_audio_transcription: Option<TranscriptionConfig>

Whether to include input audio transcription.

temperature: Option<f32>

Temperature for response generation.

max_response_output_tokens: Option<u32>

Maximum output tokens.

cached_content: Option<String>

Cached content resource name (e.g. cachedContents/1234).

interruption_detection: Option<InterruptionDetection>

Interruption detection mode for voice sessions.

Controls whether the session automatically detects user speech to cancel agent audio output, or relies on explicit API calls. Defaults to Manual when None.

affective_dialog: Option<bool>

Enable emotion-aware (“affective”) dialog, where the model adapts its tone to the user’s emotional state. Gemini Live native-audio models only (requires the v1alpha endpoint); ignored by other providers.

extra: Option<Value>

Provider-specific options.

Implementations

impl RealtimeConfig

pub fn new() -> RealtimeConfig

Create a new empty configuration.

pub fn builder() -> RealtimeConfigBuilder

Create a builder for RealtimeConfig.

pub fn with_model(self, model: impl Into<String>) -> RealtimeConfig

Set the model.

pub fn with_instruction(self, instruction: impl Into<String>) -> RealtimeConfig

Set the system instruction.

pub fn with_voice(self, voice: impl Into<String>) -> RealtimeConfig

Set the voice.

pub fn with_modalities(self, modalities: Vec<String>) -> RealtimeConfig

Set output modalities.

pub fn with_text_and_audio(self) -> RealtimeConfig

Enable text and audio output.

pub fn with_audio_only(self) -> RealtimeConfig

Enable audio-only output.

pub fn with_vad(self, vad: VadConfig) -> RealtimeConfig

Set VAD configuration.

pub fn with_server_vad(self) -> RealtimeConfig

Enable server-side VAD with default settings.

pub fn without_vad(self) -> RealtimeConfig

Disable automatic VAD, taking ownership of turn detection.

See VadMode::None — on Gemini Live this is also what unlocks client activityStart / activityEnd signalling, which the protocol permits only while server-side detection is disabled.

pub fn with_tool(self, tool: ToolDefinition) -> RealtimeConfig

Add a tool definition.

pub fn with_tools(self, tools: Vec<ToolDefinition>) -> RealtimeConfig

Set multiple tools.

pub fn with_transcription(self) -> RealtimeConfig

Enable input audio transcription.

pub fn with_temperature(self, temp: f32) -> RealtimeConfig

Set temperature.

pub fn with_affective_dialog(self, enabled: bool) -> RealtimeConfig

Enable emotion-aware (“affective”) dialog. Only honored by Gemini Live native-audio models (v1alpha); a no-op for other providers/models.

pub fn with_cached_content(self, content: impl Into<String>) -> RealtimeConfig

Set cached content resource.

pub fn with_interruption_detection( self, mode: InterruptionDetection, ) -> RealtimeConfig

Set the interruption detection mode.

See InterruptionDetection for details on each variant.

pub fn with_automatic_interruption(self) -> RealtimeConfig

Enable automatic interruption detection.

The session will detect user speech onset and cancel the current agent audio output automatically.

Trait Implementations

impl Clone for RealtimeConfig

fn clone(&self) -> RealtimeConfig

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for RealtimeConfig

fn fmt(&self, f: &mut Formatter<'_>) -> Result<(), Error>

Formats the value using the given formatter.

impl Default for RealtimeConfig

fn default() -> RealtimeConfig

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for RealtimeConfig

fn deserialize<__D>( __deserializer: __D, ) -> Result<RealtimeConfig, <__D as Deserializer<'de>>::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl From<RealtimeConfig> for SessionUpdateConfig

fn from(config: RealtimeConfig) -> SessionUpdateConfig

Converts to this type from the input type.

impl PartialEq for RealtimeConfig

fn eq(&self, other: &RealtimeConfig) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl Serialize for RealtimeConfig

fn serialize<__S>( &self, __serializer: __S, ) -> Result<<__S as Serializer>::Ok, <__S as Serializer>::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl StructuralPartialEq for RealtimeConfig