Building Agents with LangGraph Course #7: Building a Multi-Agent System LangGraph

Welcome to the seventh part of our ongoing series“Building Agents with LangGraph” course! So far, we’ve explored agents with a single Large Language Model (LLM) and a simple state. Now, we’re going to create a much more sophisticated agent composed of multiple, distinct LLM calls and a more complex state.

In this tutorial, we will build a multi-agent system that collaborates to write an essay. This system will follow a cyclical, reflective process: it will plan, research, write, and then critique its own work to produce a refined final draft.

This demonstrates how LangGraph can be used to orchestrate complex, stateful workflows involving multiple specialized “agents” or roles.

Table of Contents:

1. Project Setup and Dependencies

2. Defining the Agent’s State

3. Creating the Agent Roles (Nodes)

• The Planner Agent

• The Writer Agent

• The Reflector Agent

• The Researcher & Critique Agent

4. Implementing the Node Functions

5. Defining the Graph’s Logic and Edges

6. Visualizing and Running the Agent

This article is the Seventh Article in the ongoing series of Building LLM Agents with LangGraph:

• Introduction to Agents & LangGraph (Published!)

• Building Simple ReAct Agent from Scratch (Published!)

• Main Building Units of LangGraph (Published!)

• Agentic Search Tools in LangGraph (Published!)

• Persistence and Streaming in LangGraph (Published!)

• Human in the Loop in LLM Agents (Published!)

• Putting it All Together! Building Essay Writer Agent (You are here!)

This series is designed to take readers from foundational knowledge to advanced practices in building LLM agents with LangGraph.

Each article delves into essential components, such as constructing simple ReAct agents from scratch, leveraging LangGraph’s building units, utilizing agentic search tools, implementing persistence and streaming capabilities, integrating human-in-the-loop interactions, and culminating in the creation of a fully functional essay-writing agent.

By the end of this series, you will have a comprehensive understanding of LangGraph, practical skills to design and deploy LLM agents, and the confidence to build customized AI-driven workflows tailored to diverse applications.

---

---

1. Project Setup and Dependencies

First, let’s set up our environment. We’ll load our API keys from a .env file and import the necessary libraries. This includes components from langgraph for building the graph, langchain_core for message types, and langchain_openai for the model. We’ll also set up an in-memory checkpointer using SqliteSaver to persist the state of our graph.

from dotenv import load_dotenv

_ = load_dotenv()

from langgraph.graph import StateGraph, END
from typing import TypedDict, Annotated, List
import operator
from langgraph.checkpoint.sqlite import SqliteSaver
from langchain_core.messages import AnyMessage, SystemMessage, HumanMessage, AIMessage, ChatMessage

memory = SqliteSaver.from_conn_string(":memory:")

---

2. Defining the Agent’s State

The state is the backbone of our agent, defining the structure of data that gets passed between nodes. For our essay writer, the state needs to track several pieces of information: the initial task, the plan, the research content, the draft, critiques, and the number of revisions.

We define this using a TypedDict:

class AgentState(TypedDict):
    task: str
    plan: str
    draft: str
    critique: str
    content: List[str]
    revision_number: int
    max_revisions: int

Next, we initialize our model. We will use GPT-4 for this example.

from langchain_openai import ChatOpenAI
model = ChatOpenAI(model="gpt-4", temperature=0)

---

3. Creating the Agent Roles (Nodes)

Our multi-agent system will consist of several specialized roles, each represented by a node in our graph. Each role has a specific prompt that guides the LLM to perform its task.

The Planner Agent

The first step is to create a high-level plan. The planner node takes the user’s task and generates an essay outline.

PLAN_PROMPT = """You are an expert writer tasked with writing a high level outline of an essay. \
Write such an outline for the user provided topic. Give an outline of the essay along with any relevant notes \
or instructions for the sections."""

The Writer Agent

The writer node is responsible for generating the essay draft based on the plan and any research content. It’s also designed to revise its draft based on critiques.