Basic LangGraph agent for stateful workflows

This example demonstrates a fundamental application of LangGraph, an extension of LangChain designed for building stateful, multi-actor applications with LLMs. It illustrates how to define a simple computational graph with a single LLM node and explicit state, allowing the workflow to keep the model output visible without redundant end-node processing.

Requirements

• Python 3.10+
• An OpenAI API key set as an environment variable (OPENAI_API_KEY)

Steps for running this example in the shell

1. Install dependencies:

python -m venv .venv

# macOS/Linux:
source .venv/bin/activate

# Windows Command Prompt:
.venv\Scripts\activate.bat

# Windows PowerShell:
.venv\Scripts\Activate.ps1

pip install -r requirements.txt

2. Export your API key as an environment variable:

export OPENAI_API_KEY="sk-..." # Windows cmd: set OPENAI_API_KEY="sk-..." # Windows PowerShell: $env:OPENAI_API_KEY="sk-..."

3. Run the script:

python langgraph_basic_agent.py

Output

When you run the script, it will execute a simple LangGraph workflow twice with different initial inputs. You will see a print statement indicating the LLM node (---CALL_LLM---), followed by the final state, which includes the LLM's response. This demonstrates how the output field in the GraphState is updated and carried through the graph.

Example output:

---CALL_LLM---
Initial Input: Hello, how are you today?
Final Output: I am an AI, so I don't have feelings, but I'm ready to assist you! How can I help you today?

---CALL_LLM---
Initial Input 2: What is the capital of France?
Final Output 2: The capital of France is Paris.