Interface ResponsesDelegationOptions

Settings for the backend Responses model. Unset fields use the service defaults.

interface ResponsesDelegationOptions {
    instructions?: string;
    maxOutputTokens?: number;
    model?: string;
    parallelToolCalls?: boolean;
    reasoning?: Reasoning;
    serviceTier?: "auto" | "default" | "flex" | "priority";
    text?: ResponseTextConfig;
    toolChoice?: null | llm.ToolChoice;
}

Properties

instructions? maxOutputTokens? model? parallelToolCalls? reasoning? serviceTier? text? toolChoice?

Properties

Optional instructions

instructions?: string

Backend instructions, separate from the voice persona.

Optional maxOutputTokens

maxOutputTokens?: number

Maximum tokens per backend response; the service requires at least 16.

Optional model

model?: string

Defaults to gpt-5.6-luna.

Optional parallelToolCalls

parallelToolCalls?: boolean

Allow the backend to request multiple tools in one response.

Optional reasoning

reasoning?: Reasoning

Reasoning effort for the backend model.

Optional serviceTier

serviceTier?: "auto" | "default" | "flex" | "priority"

Processing tier for backend responses.

Optional text

text?: ResponseTextConfig

Text format and verbosity for backend responses.

Optional toolChoice

toolChoice?: null | llm.ToolChoice

Restrict backend tool selection. Null restores automatic selection.

Settings

Member Visibility

Theme

On This Page

instructionsmaxOutputTokensmodelparallelToolCallsreasoningserviceTiertexttoolChoice