Interface AgentExecutorInput

Interface defining the structure of input data for creating an AgentExecutor. It extends ChainInputs and includes additional properties specific to agent execution.

Hierarchy

Properties

agent tools callbackManager? callbacks? earlyStoppingMethod? handleParsingErrors? maxIterations? memory? metadata? returnIntermediateSteps? tags? verbose?

Properties

agent

agent: BaseSingleActionAgent | BaseMultiActionAgent | Runnable<ChainValues & {
    steps?: AgentStep[];
}, AgentAction | AgentFinish | AgentAction[], BaseCallbackConfig>

tools

tools: (StructuredTool<ZodObject<any, any, any, any, {}>> | StructuredTool<ZodObject<any, any, any, any, {}>>)[]

Optional callbackManager

callbackManager?: CallbackManager

Deprecated

Use callbacks instead

Optional callbacks

callbacks?: Callbacks

Optional earlyStoppingMethod

earlyStoppingMethod?: StoppingMethod

Optional handleParsingErrors

handleParsingErrors?: string | boolean | ((e) => string)

Type declaration

Optional maxIterations

maxIterations?: number

Optional memory

memory?: BaseMemory

Optional metadata

metadata?: Record<string, unknown>

Optional returnIntermediateSteps

returnIntermediateSteps?: boolean

Optional tags

tags?: string[]

Optional verbose

verbose?: boolean

Settings

Member Visibility

Theme

On This Page

Generated using TypeDoc