LangChain Reference home page

GitHub
Main Docs

MCP Adapters

Standard Tests

Text Splitters

Overview
Agents
Callbacks
Chains
Chat models
Embeddings
Evaluation
Globals
Hub
Memory
Output parsers
Retrievers
Runnables
LangSmith
Storage

⌘I

LangChain Assistant

Ask a question to get started

Enter to send•Shift+Enter new line

Menu

Overview Agents Callbacks Chains Chat models Embeddings Evaluation Globals Hub Memory Output parsers Retrievers Runnables LangSmith Storage

Language

Theme

Pythonlangchain-classicagentsagentAgent

Class●Since v1.0Deprecated

Agent

Copy

Agent()

Bases

BaseSingleActionAgent

Attributes

Methods

Inherited fromBaseSingleActionAgent

Methods

Save the agent.

View source on GitHub

llm_chain: LLMChain

LLMChain to use for agent.

output_parser: AgentOutputParser

Output parser to use for agent.

allowed_tools: list[str] | None

Allowed tools for the agent. If None, all tools are allowed.

return_values: list[str]

Return values of the agent.

input_keys: list[str]

Return the input keys.

observation_prefix: str

Prefix to append the observation with.

llm_prefix: str

Prefix to append the LLM call with.

Return dictionary representation of agent.

get_allowed_tools

Get allowed tools.

Given input, decided what to do.

Async given input, decided what to do.

get_full_inputs

Create the full inputs for the LLMChain from intermediate steps.

validate_prompt

Validate that prompt matches format.

Create a prompt for this class.

from_llm_and_tools

Construct an agent from an LLM and tools.

return_stopped_response

Return response when agent has been stopped due to max iterations.

tool_run_logging_kwargs

Return logging kwargs for tool run.

Agent that calls the language model and deciding the action.

This is driven by a LLMChain. The prompt in the LLMChain MUST include a variable called "agent_scratchpad" where the agent can put its intermediary work.