Member Junction
    Preparing search index...

    Interface RealtimeRemoteReasoning

    Configuration for remote reasoning delegation.

    interface RealtimeRemoteReasoning {
        Effort?: "none" | "low" | "medium" | "high" | "minimal" | "xhigh";
        Kind?: "model" | "hostedAgent";
        MaxOutputTokens?: number;
        Ref?: string;
    }
    Index

    Properties

    Effort? Kind? MaxOutputTokens? Ref?

    Properties

    Effort?: "none" | "low" | "medium" | "high" | "minimal" | "xhigh"

    Reasoning effort level for supported models.

    Kind?: "model" | "hostedAgent"

    What the remote reference denotes. model = Live/Inworld; hostedAgent = ElevenLabs.

    MaxOutputTokens?: number

    Maximum output tokens for the remote reasoning pass.

    Ref?: string

    'gpt-5.6-terra' | 'anthropic/claude-sonnet-4-6' | 'MJ Realtime Co-Agent'.

    Settings

    Member Visibility

    On This Page

    Properties