All Classes and Interfaces
Class
Description
Marks a method as an A2A client agent.
Marks a method as an activation predicate for one or more sub-agents of a conditional agent.
Java methods annotated with
@Agent are considered agents that other agents can invoke.The AgenticScope class represents a common environment where agents belonging to the same
agentic system can share their state.
Allow to access the
AgenticScope of any agent extending it.A codec for serializing and deserializing
DefaultAgenticScope objects to and from JSON.Singleton registry for managing AgenticScope instances.
Utility class for serializing AgenticScope objects to JSON format.
Service Provider Interface for AgenticScope persistence.
Provides static factory methods to create and configure various types of agent services.
Listener interface for monitoring agent invocations.
Marks a method as a supplier of a listener for this agent.
Monitors agent executions and provides observability for the LangChain4j Agentic system.
Marks a method as a supplier of the
ChatMemoryProvider that an agent can utilize during its operation.Marks a method as a supplier of chat memory that an agent can utilize during its operation.
Marks a method as a supplier of the chat model to be used by an agent.
Marks a method as a definition of a conditional agent, generally used to route the agentic workflow toward
one or more sub-agents according to the verification of their activation conditions.
Marks a method as a supplier of content retriever that an agent can utilize during its operation.
Marks a method as an error handler for a workflow agent.
Marks a method as an exit predicate for a loop in a loop-based agent.
Generates HTML reports for agent executions.
Marks a method as a human-in-the-loop agent.
A parameter of an agentic method is annotated with
@K to indicate that it is populated with
the value of a key representing a specific typed state in the agentic system.Marks a method as a definition of a loop agent, used to orchestrate the agentic workflow
by invoking a series of sub-agents in a loop until a certain predicate is met or a maximum number of iterations is reached.
Marks a parameter that will receive the current loop iteration count in a loop-based agent.
Wraps an existing
AgentInvoker to inject a specific item from a collection
into the agent's invocation arguments.Marks a method as an MCP client agent that wraps a single MCP tool as a non-AI agent.
Marks a static method that returns the McpClient instance for declarative MCP client agents.
Interface that agent service interfaces can extend to automatically register
an
AgentMonitor as a listener.Represents a monitored execution of an agentic system, tracking the top-level agent invocation
and any other nested invocations, along with all the invocation currently in progress and
any errors that occur during execution.
Marks a method as the output definition of a workflow agent,
generally combining results from different states of the
AgenticScope.Marks a method as a definition of a parallel agent, used to orchestrate the agentic workflow
by invoking a series of sub-agents in parallel.
Used in combination with
ParallelAgent to specify the executor that will be used to run the sub-agents in parallel.Marks a method as a definition of a parallel multi-instance agent, used to orchestrate the agentic workflow
by creating multiple instances of the same sub-agent, one for each item in a collection read from the agentic scope.
A
DelayedResponse that can be completed externally, without spawning a background thread.Marks a method as a definition of an agent based on a custom planner.
Marks a method as a supplier of the planner for a planner based agent.
Holds the result of an agent invocation along with its associated
AgenticScope.Marks a method as a supplier of a retrieval augmentor that an agent can utilize during its operation.
Marks a method as a definition of a sequence agent, used to orchestrate the agentic workflow
by invoking a series of sub-agents in a predefined order.
Marks a method as a supplier of the streaming chat model to be used by an agent.
Marks a method as a supervisor agent that can autonomously coordinate and invoke multiple sub-agents.
Strategy for providing context to the supervisor agent.
Marks a method as a definition of the request that will be sent to the supervisor.
Strategy to decide which response the supervisor agent should return.
Marks a method as a supplier of a tool provider that an agent can utilize during its operation.
Marks a method as a supplier of tools that an agent can utilize during its operation.
A class implementing this interface represents the input or output key of an agent in a strongly typed way.