Class AgentExecutor

A chain managing an agent using tools.

Example


const executor = AgentExecutor.fromAgentAndTools({
  agent: async () => loadAgentFromLangchainHub(),
  tools: [new SerpAPI(), new Calculator()],
  returnIntermediateSteps: true,
});

const result = await executor.invoke({
  input: `Who is Olivia Wilde's boyfriend? What is his current age raised to the 0.23 power?`,
});

Hierarchy (View Summary)

BaseChain<Toolkit, AgentExecutorOutput>
- AgentExecutor

Constructors

constructor

new AgentExecutor(input: AgentExecutorInput): AgentExecutor
Parameters
- input: AgentExecutorInput
Returns AgentExecutor
Overrides BaseChain.constructor
- Defined in remotes/langchain-ai/langchainjs/main/libs/langchain-classic/src/agents/executor.ts:400

Properties

agent

agent: BaseSingleActionAgent | BaseMultiActionAgent

earlyStoppingMethod

earlyStoppingMethod: StoppingMethod = "force"

handleParsingErrors

handleParsingErrors: string | boolean | ((e: any) => string) = false

How to handle errors raised by the agent's output parser. Defaults to False, which raises the error.

If true, the error will be sent back to the LLM as an observation. If a string, the string itself will be sent to the LLM as an observation. If a callable function, the function will be called with the exception as an argument, and the result of that function will be passed to the agent as an observation.

`Optional`handleToolRuntimeErrors

handleToolRuntimeErrors?: (e: Error) => string

`Optional`maxIterations

maxIterations?: number = 15

`Optional`memory

memory?: any

returnIntermediateSteps

returnIntermediateSteps: boolean = false

returnOnlyOutputs

returnOnlyOutputs: boolean = true

tools

tools: StructuredToolInterface[]

Accessors

inputKeys

get inputKeys(): string[]
Returns string[]
Overrides BaseChain.inputKeys
- Defined in remotes/langchain-ai/langchainjs/main/libs/langchain-classic/src/agents/executor.ts:392

lc_namespace

get lc_namespace(): string[]
Returns string[]
Overrides BaseChain.lc_namespace
- Defined in remotes/langchain-ai/langchainjs/main/libs/langchain-classic/src/agents/executor.ts:357

outputKeys

get outputKeys(): string[]
Returns string[]
Overrides BaseChain.outputKeys
- Defined in remotes/langchain-ai/langchainjs/main/libs/langchain-classic/src/agents/executor.ts:396

shouldContinueGetter

get shouldContinueGetter(): (iterations: number) => boolean
Returns (iterations: number) => boolean
- - (iterations: number): boolean
  - Method that checks if the agent execution should continue based on the number of iterations.
    Parameters
    iterations: number
    The current number of iterations.
    Returns boolean
    A boolean indicating whether the agent execution should continue.
- Defined in remotes/langchain-ai/langchainjs/main/libs/langchain-classic/src/agents/executor.ts:456

Methods

_chainType

_chainType(): "agent_executor"
Return the string type key uniquely identifying this class of chain.

Returns "agent_executor"
Overrides BaseChain._chainType
- Defined in remotes/langchain-ai/langchainjs/main/libs/langchain-classic/src/agents/executor.ts:801

`Protected`_formatValues

_formatValues(values: any): Promise<any>
Parameters
- values: any
Returns Promise<any>
Inherited from BaseChain._formatValues
- Defined in remotes/langchain-ai/langchainjs/main/libs/langchain-classic/src/chains/base.ts:217

_getToolReturn

_getToolReturn(nextStepOutput: AgentStep): Promise<any>
Parameters
- nextStepOutput: AgentStep
Returns Promise<any>
- Defined in remotes/langchain-ai/langchainjs/main/libs/langchain-classic/src/agents/executor.ts:746

_return

_return(
    output: AgentFinish,
    intermediateSteps: AgentStep[],
    runManager?: any,
): Promise<ChainValues>
Parameters
- output: AgentFinish
- intermediateSteps: AgentStep[]
- OptionalrunManager: any
Returns Promise<ChainValues>
- Defined in remotes/langchain-ai/langchainjs/main/libs/langchain-classic/src/agents/executor.ts:731

_returnStoppedResponse

_returnStoppedResponse(earlyStoppingMethod: StoppingMethod): AgentFinish
Parameters
- earlyStoppingMethod: StoppingMethod
Returns AgentFinish
- Defined in remotes/langchain-ai/langchainjs/main/libs/langchain-classic/src/agents/executor.ts:764

_streamIterator

_streamIterator(
inputs: Record<string, any>,
options?: any,
): AsyncGenerator<ChainValues>
Parameters
- inputs: Record<string, any>
- Optionaloptions: any
Returns AsyncGenerator<ChainValues>
- Defined in remotes/langchain-ai/langchainjs/main/libs/langchain-classic/src/agents/executor.ts:778

_takeNextStep

_takeNextStep(
    nameToolMap: Record<string, Toolkit>,
    inputs: ChainValues,
    intermediateSteps: AgentStep[],
    runManager?: any,
    config?: any,
): Promise<any>
Parameters
- nameToolMap: Record<string, Toolkit>
- inputs: ChainValues
- intermediateSteps: AgentStep[]
- OptionalrunManager: any
- Optionalconfig: any
Returns Promise<any>
- Defined in remotes/langchain-ai/langchainjs/main/libs/langchain-classic/src/agents/executor.ts:624

apply

apply(inputs: ChainValues[], config?: any[]): Promise<ChainValues[]>
Parameters
- inputs: ChainValues[]
- Optionalconfig: any[]
Returns Promise<ChainValues[]>
Deprecated
Use .batch() instead. Will be removed in 0.2.0.

Call the chain on all inputs in the list
Inherited from BaseChain.apply
- Defined in remotes/langchain-ai/langchainjs/main/libs/langchain-classic/src/chains/base.ts:258

call

call(values: any, config?: any, tags?: string[]): Promise<ChainValues>
Parameters
- values: any
- Optionalconfig: any
- Optionaltags: string[]
  Deprecated
Returns Promise<ChainValues>
Deprecated
Use .invoke() instead. Will be removed in 0.2.0.

Run the core logic of this chain and add to output if desired.

Wraps _call and handles memory.
Inherited from BaseChain.call
- Defined in remotes/langchain-ai/langchainjs/main/libs/langchain-classic/src/chains/base.ts:243

invoke

invoke(input: ChainValues, options?: any): Promise<ChainValues>
Invoke the chain with the provided input and returns the output.
Parameters
- input: ChainValues
  Input values for the chain run.
- Optionaloptions: any
Returns Promise<ChainValues>
Promise that resolves with the output of the chain run.
Inherited from BaseChain.invoke
- Defined in remotes/langchain-ai/langchainjs/main/libs/langchain-classic/src/chains/base.ts:86

prepOutputs

prepOutputs(
    inputs: Record<string, unknown>,
    outputs: Record<string, unknown>,
    returnOnlyOutputs?: boolean,
): Promise<Record<string, unknown>>
Parameters
- inputs: Record<string, unknown>
- outputs: Record<string, unknown>
- returnOnlyOutputs: boolean = false
Returns Promise<Record<string, unknown>>
Inherited from BaseChain.prepOutputs
- Defined in remotes/langchain-ai/langchainjs/main/libs/langchain-classic/src/chains/base.ts:149

run

run(input: any, config?: any): Promise<string>
Parameters
- input: any
- Optionalconfig: any
Returns Promise<string>
Deprecated
Use .invoke() instead. Will be removed in 0.2.0.
Inherited from BaseChain.run
- Defined in remotes/langchain-ai/langchainjs/main/libs/langchain-classic/src/chains/base.ts:190

serialize

serialize(): SerializedLLMChain
Return a json-like object representing this chain.

Returns SerializedLLMChain
Overrides BaseChain.serialize
- Defined in remotes/langchain-ai/langchainjs/main/libs/langchain-classic/src/agents/executor.ts:805

`Static`deserialize

deserialize(
data: SerializedBaseChain,
values?: LoadValues,
): Promise<BaseChain<ChainValues, ChainValues>>
Load a chain from a json-like object describing it.
Parameters
- data: SerializedBaseChain
- values: LoadValues = {}
Returns Promise<BaseChain<ChainValues, ChainValues>>
Inherited from BaseChain.deserialize
- Defined in remotes/langchain-ai/langchainjs/main/libs/langchain-classic/src/chains/base.ts:270

`Static`fromAgentAndTools

fromAgentAndTools(fields: AgentExecutorInput): AgentExecutor
Create from agent and a list of tools.
Parameters
- fields: AgentExecutorInput
Returns AgentExecutor
- Defined in remotes/langchain-ai/langchainjs/main/libs/langchain-classic/src/agents/executor.ts:452

`Static`lc_name

lc_name(): string
Returns string
- Defined in remotes/langchain-ai/langchainjs/main/libs/langchain-classic/src/agents/executor.ts:353

Class AgentExecutor

Example

Hierarchy (View Summary)

Index

Constructors

Properties

Accessors

Methods

Constructors

constructor

Parameters

Returns AgentExecutor

Properties

agent

earlyStoppingMethod

handleParsingErrors

OptionalhandleToolRuntimeErrors

OptionalmaxIterations

Optionalmemory

returnIntermediateSteps

returnOnlyOutputs

tools

Accessors

inputKeys

Returns string[]

lc_namespace

Returns string[]

outputKeys

Returns string[]

shouldContinueGetter

Returns (iterations: number) => boolean

Parameters

Returns boolean

Methods

_chainType

Returns "agent_executor"

Protected_formatValues

Parameters

Returns Promise<any>

_getToolReturn

Parameters

Returns Promise<any>

_return

Parameters

Returns Promise<ChainValues>

_returnStoppedResponse

Parameters

Returns AgentFinish

_streamIterator

Parameters

Returns AsyncGenerator<ChainValues>

_takeNextStep

Parameters

Returns Promise<any>

apply

Parameters

Returns Promise<ChainValues[]>

Deprecated

call

Parameters

Deprecated

Returns Promise<ChainValues>

Deprecated

invoke

Parameters

Returns Promise<ChainValues>

prepOutputs

Parameters

Returns Promise<Record<string, unknown>>

run

Parameters

Returns Promise<string>

Deprecated

serialize

Returns SerializedLLMChain

Staticdeserialize

Parameters

Returns Promise<BaseChain<ChainValues, ChainValues>>

StaticfromAgentAndTools

Parameters

`Optional`handleToolRuntimeErrors

`Optional`maxIterations

`Optional`memory

`Protected`_formatValues

`Static`deserialize

`Static`fromAgentAndTools

`Static`lc_name