Kenhuang Academy
  • English

      • Quick Start

      This chapter builds your first tool-using agent with LangChain 1.0's create_agent.

      Minimal agent

      from langchain.agents import create_agent

def get_weather(city: str) -> str:
    """Return fake weather for demo."""
    return f"It's sunny in {city}."

agent = create_agent(
    "openai:gpt-4.1-mini",
    tools=[get_weather],
    system_prompt="You are a helpful assistant. Use tools when needed.",
)

result = agent.invoke({
    "messages": [{"role": "user", "content": "What's the weather in Shanghai?"}]
})

print(result["messages"][-1].content)

      What happens:

      1. The model reads the user message
      2. It chooses to call get_weather
      3. The framework runs the tool and feeds the result back
      4. The model writes the final answer

      @tool decorator

      Add types and docstrings so the model knows how to call tools:

      from langchain.tools import tool
from langchain.agents import create_agent

@tool
def search(query: str) -> str:
    """Search the web for information."""
    return f"Results for: {query}"

agent = create_agent("openai:gpt-4.1-mini", tools=[search])

      Streaming

      for chunk in agent.stream(
    {"messages": [{"role": "user", "content": "Explain RAG in one paragraph."}]},
    stream_mode="messages",
):
    if chunk.content:
        print(chunk.content, end="", flush=True)

      Use stream_mode="updates" to see per-node updates (including tool calls).

      Structured output

      from pydantic import BaseModel
from langchain.agents import create_agent

class Answer(BaseModel):
    summary: str
    confidence: float

agent = create_agent(
    "openai:gpt-4.1-mini",
    tools=[],
    response_format=Answer,
)

result = agent.invoke({
    "messages": [{"role": "user", "content": "Summarize LangChain 1.0 in one sentence."}]
})
print(result["structured_response"])

      Multi-turn memory (preview)

      Use a checkpointer and thread_id (Memory & Checkpoints):

      from langchain.agents import create_agent
from langgraph.checkpoint.memory import InMemorySaver

agent = create_agent(
    "openai:gpt-4.1-mini",
    tools=[],
    checkpointer=InMemorySaver(),
)

config = {"configurable": {"thread_id": "demo-thread-1"}}

agent.invoke(
    {"messages": [{"role": "user", "content": "My name is Alex."}]},
    config=config,
)
result = agent.invoke(
    {"messages": [{"role": "user", "content": "What's my name?"}]},
    config=config,
)
print(result["messages"][-1].content)

      Checklist

      1. venv + pip install langchain langgraph langchain-openai
      2. Set OPENAI_API_KEY
      3. Run the minimal agent
      4. Add a @tool and watch tool calls
      5. Try stream and checkpointer

      Next steps