react_agent

ReAct Agent (Functional API)

Demonstrates the ReAct agent pattern (think -> act -> observe -> repeat) as a natural while loop using LangGraph's @task and @entrypoint decorators.

What This Sample Demonstrates

• The ReAct loop as a natural while True loop in Python
• @task functions for agent thinking and tool execution
• How the Functional API makes agent loops readable and extensible

How It Works

1. The agent_think task examines the query and tool history, deciding the next action.
2. If a tool is needed, execute_tool runs it and the result is appended to history.
3. The loop continues until agent_think returns a final answer.

while True:
    decision = await agent_think(query, history)
    if decision["action"] == "final":
        return decision["answer"]
    result = await execute_tool(decision["tool_name"], decision["tool_input"])
    history.append(result)

Running the Sample

Prerequisites: uv sync --group langgraph and a running Temporal dev server (temporal server start-dev).

# Terminal 1
uv run langgraph_plugin/functional_api/react_agent/run_worker.py

# Terminal 2
uv run langgraph_plugin/functional_api/react_agent/run_workflow.py

Files

File	Description
workflow.py	@task functions (agent_think, execute_tool), @entrypoint, and ReactAgentFunctionalWorkflow
run_worker.py	Registers tasks and entrypoint with LangGraphPlugin, starts worker
run_workflow.py	Executes the agent workflow and prints the answer