Skip to main content
Agent components define the behavior and capabilities of AI agents in your flow. Agents use LLMs as a reasoning engine to decide which of the connected tool components to use to solve a problem. Tools in agentic functions are essentially functions that the agent can call to perform tasks or access external resources. A function is wrapped as a Tool object with a common interface that the agent understands. Agents become aware of tools through tool registration, where the agent is provided a list of available tools typically at agent initialization. The Tool object’s description tells the agent what the tool can do. The agent then uses a connected LLM to reason through the problem to decide which tool is best for the job.

Use an agent in a flow

The simple agent starter project uses an agent component connected to a ,URL and Calculator tools to answer a user’s questions. The OpenAI LLM acts as a brain for the agent to decide which tool to use. Tools are connected to agent components at the Tools port. Agent In A Flow Pn

Agent component

This component creates an agent that can use tools to answer questions and perform tasks based on given instructions. The component includes an LLM model integration, a system message prompt, and a Tools port to connect tools to extend its capabilities. For more information on this component,
NameTypeDescription
agent_llmDropdownThe provider of the language model that the agent uses to generate responses. Options include OpenAI and other providers or Custom.
system_promptStringThe system prompt provides initial instructions and context to guide the agent’s behavior.
toolsListThe list of tools available for the agent to use. This field is optional and can be empty.
input_valueStringThe input task or question for the agent to process.
add_current_date_toolBooleanWhen true this adds a tool to the agent that returns the current date.
memoryMemoryAn optional memory configuration for maintaining conversation history.
max_iterationsIntegerThe maximum number of iterations the agent can perform.
handle_parsing_errorsBooleanThis determines whether to handle parsing errors during agent execution.
verboseBooleanThis enables verbose output for detailed logging.
Outputs
NameTypeDescription
responseMessageThe agent’s response to the given input task.

JSON Agent

This component creates a JSON agent from a JSON or YAML file and an LLM.
NameTypeDescription
llmLanguageModelThe language model to use for the agent.
pathFileThe path to the JSON or YAML file.
Outputs
NameTypeDescription
agentAgentExecutorThe JSON agent instance.

Vector Store Agent

This component creates a Vector Store Agent using LangChain.
Inputs
NameTypeDescription
llmLanguageModelThe language model to use for the agent.
vectorstoreVectorStoreInfoThe vector store information for the agent to use.
Outputs
NameTypeDescription
agentAgentExecutorThe Vector Store Agent instance.

Vector Store Router Agent

This component creates a Vector Store Router Agent using LangChain.
Inputs
NameTypeDescription
llmLanguageModelThe language model to use for the agent.
vectorstoresList[VectorStoreInfo]The list of vector store information for the agent to route between.
Outputs
NameTypeDescription
agentAgentExecutorThe Vector Store Router Agent instance.