ART Framework API Docs
    Preparing search index...

    Interface AgentOptions

    Options to override agent behavior at runtime.

    AgentOptions

    interface AgentOptions {
        executionConfig?: ExecutionConfig;
        forceTools?: string[];
        llmParams?: Record<string, any>;
        overrideModel?: { model: string; provider: string };
        persona?: Partial<AgentPersona>;
        promptTemplateId?: string;
        providerConfig?: RuntimeProviderConfig;
        stream?: boolean;
        systemPrompt?: string | SystemPromptOverride;
    }
    Index

    Properties

    executionConfig?: ExecutionConfig

    Optional: Configuration for execution phase behavior (TAEF parameters) for this specific call. Overrides thread and instance-level execution config.

    forceTools?: string[]

    Force the use of specific tools, potentially overriding the thread's enabledTools for this call (use with caution).

    llmParams?: Record<string, any>

    Override specific LLM parameters (e.g., temperature, max_tokens) for this call only.

    overrideModel?: { model: string; provider: string }

    Specify a particular reasoning model to use for this call, overriding the thread's default.

    persona?: Partial<AgentPersona>

    Optional: Defines the identity and high-level guidance for the agent for this specific call. This overrides both the instance-level and thread-level persona.

    promptTemplateId?: string

    Override the prompt template used for this specific call.

    providerConfig?: RuntimeProviderConfig

    Override provider configuration for this specific call.

    stream?: boolean

    Request a streaming response for this specific agent process call.

    systemPrompt?: string | SystemPromptOverride

    Optional system prompt override/tag to override thread, instance, or agent defaults for this specific call.