DEV Community

Rushank Savant
Rushank Savant

Posted on

Day 2: Setting Up & Your First "Hello World" Chain 🛠️

#ai #langchain #python #tutorial

Welcome back! Yesterday, we talked about the "what" and "why." Today, we stop talking and start building.

Official documentations can sometimes feel like a maze of installation commands. Let’s cut through the noise and get your environment ready so you can run your first "Hello World" AI chain in under 5 minutes.

🛠️ Step 1: The Toolbox (Installation)

First, we need the right tools. Open your terminal and create a new folder for this project. We’ll use a virtual environment to keep things clean.

If you are interested to learn in a bit more detail about virtual environment, visit my Python Environment blog. Here I talk about uv — the modern venv tool.

# Create and activate your environment
python -m venv langchain-env

source langchain-env/bin/activate  # On Windows use: .\langchain-env\Scripts\activate

# Install the essential packages
pip install langchain langchain-openai python-dotenv
Enter fullscreen mode Exit fullscreen mode

langchain: The core framework.

langchain-openai: The specific "plug-in" to talk to OpenAI models (modern LangChain is modular!).

python-dotenv: A best-practice tool to keep your API keys secret.

🔑 Step 2: The Secret Key

Never hardcode your API keys in your code! It's a habit that will save you from accidental leaks later.

Create a file named .env in your project folder.

Add your key like this:
OPENAI_API_KEY=your_actual_key_here

🏗️ Step 3: Building the "Hello World" Chain

In the modern LangChain era (2026), we use LCEL (LangChain Expression Language). It uses a "Pipe" operator (|) that makes code look like a clean flow chart.

Create a file named app.py and paste this in:

import os
from dotenv import load_dotenv
from langchain_openai import ChatOpenAI
from langchain_core.prompts import ChatPromptTemplate
from langchain_core.output_parsers import StrOutputParser

# 1. Load the API key
load_dotenv()

# 2. Initialize the "Brain" (The Model)
model = ChatOpenAI(model="gpt-4o-mini")

# 3. Create the "Instructions" (The Prompt)
prompt = ChatPromptTemplate.from_template("Tell me a fun fact about {topic}")

# 4. Create the "Cleaner" (The Output Parser)
parser = StrOutputParser()

# 5. The Magic: The Pipe Operator!
# This "chains" them together: Prompt -> Model -> Parser
chain = prompt | model | parser

# 6. Run it!
response = chain.invoke({"topic": "space"})
print(response)
Enter fullscreen mode Exit fullscreen mode

🤔 What just happened? (The Breakdown)

Instead of a messy block of code, we built a pipeline:

The Prompt: We told the AI how to act. The {topic} is a placeholder we fill in later.

The Model: We connected to the LLM brain.

The Parser: By default, AI models return a complex object with metadata. The StrOutputParser simply "snips" out just the text for us.

The Pipe (|): This is the heart of LangChain. It takes the output of one step and feeds it directly into the next.

🎯 Day 2 Summary

You successfully:

  • Set up a professional dev environment.
  • Secured your credentials.
  • Built a reusable pipeline using the modern LCEL syntax.

Your Homework: Try changing the prompt! Instead of a "fun fact," tell the AI to "Write a 2-line poem about {topic} in the style of a pirate."
See how easy it is to change the behavior?

See you then! ✨

Top comments (0)