@squidcloud/client
    Preparing search index...

    Interface StartOpenAiWebVoiceCallRequest

    Request to start a voice call with an agent from a browser. The browser produces a WebRTC offer, Squid creates the call on the OpenAI Realtime API and returns the answer; from then on the browser talks to OpenAI directly while Squid runs the agent behind the call. Every setting defaults to what the agent's phone line uses (its Twilio ability with the OpenAI engine), then to the application's OpenAI Voice integration. Open to application users under the agent's security rules, and to the API key.

    interface StartOpenAiWebVoiceCallRequest {
        additionalLanguages?: string[];
        agentId: string;
        allowInterruptions?: boolean;
        chatOptions?: AllAiAgentChatOptions;
        greeting?: string;
        greetingMode?: "phrase" | "agent";
        instructions?: string;
        integrationId?: string;
        interruptionSensitivity?: "low" | "high" | "normal";
        language?: string;
        languageMode?: AiVoiceLanguageMode;
        metadata?: Record<string, unknown>;
        model?: string;
        sdp: string;
        voice?: string;
    }
    Index
    additionalLanguages? agentId allowInterruptions? chatOptions? greeting? greetingMode? instructions? integrationId? interruptionSensitivity? language? languageMode? metadata? model? sdp voice?
    additionalLanguages?: string[]

    The languages the caller may switch to in selected mode.

    agentId: string

    The agent the realtime model asks for knowledge and actions.

    allowInterruptions?: boolean

    Whether caller speech interrupts the model; defaults to the phone line, then OpenAI's default.

    chatOptions?: AllAiAgentChatOptions

    Chat options applied to every turn the agent runs for the call: instructions, agent context.

    greeting?: string

    Spoken as soon as the call starts; without any greeting the model waits for the caller.

    greetingMode?: "phrase" | "agent"

    agent has the agent compose the opening line, with greeting as its fallback; defaults to the phone line's mode.

    instructions?: string

    What this call is for, added to the realtime model's instructions.

    integrationId?: string

    The openai_voice integration to use; needed only when the agent has no phone line and the app has several.

    interruptionSensitivity?: "low" | "high" | "normal"

    How readily caller speech is detected; defaults to the phone line, then OpenAI's default.

    language?: string

    Language the call starts in, as an ISO code or BCP-47 tag.

    languageMode?: AiVoiceLanguageMode

    Whether the caller may switch language mid-call.

    metadata?: Record<string, unknown>

    Free-form data echoed on every session event, for the caller's own bookkeeping.

    model?: string

    Realtime model that talks on the call.

    sdp: string

    The browser's WebRTC offer, as SDP.

    voice?: string

    Voice of the realtime model.

    Settings
    Member Visibility

    On This Page
    Properties