Intro

Hello folks! My name is Sudharshini Jothikumar.
In this blog, we are going to explore how to create a bot without using any paid APIs — fully free, offline-friendly, and with minimal code. Whether you're a beginner or just tired of API keys and limits, this one's for you

🧠 Wait… What Are All These Fancy Words?

Sometimes we casually throw around cool tech terms like LLM, LLaMA, Ollama, LangChain, and RAG...
But do we really know what they mean? 😅
Let’s break them down in a simple, friendly way 💖

🦙 LLaMA – Meta’s Very Own LLM

LLaMA stands for "Large Language Model Meta AI"
It’s an open-source AI model created by Meta (Facebook) to understand and generate human-like text.

Think of LLaMA as a brain that was trained on lots of data from the internet — books, posts, articles, code — so it can talk like a human!

🧰 Ollama – Running LLaMA on Your Laptop

Ollama is a magical tool that lets you run LLaMA (and other models) directly on your PC or server — no internet required, no APIs, no cloud!

🧱 LangChain – Build Chatbots Like LEGO

LangChain is a Python (and JS) library designed to help you build chatbots and AI assistants more easily.

It connects:

• Your AI model (like LLaMA)

• Tools (like memory, databases)

• Logic (like chains and prompts)

  🔁 RAG – Teach Your Bot Custom Knowledge

  RAG stands for Retrieval-Augmented Generation.
  It’s a way to train your AI assistant on your own documents or data (like PDFs, medical records, books).

• 🧪 Step 1: Install Ollama

  Ollama helps you run LLMs like LLaMA on your own computer — no internet required after download. ✨

  👉 Download Ollama from here:
  🔗 https://ollama.com

  Just pick your OS (Windows, Mac, or Linux) and follow the installation guide. It’s super easy.

• 🔧 Step 2: Open Your Terminal

  After installation, open your command line/terminal and simply type:

    ollama
  

  If it runs without errors, Ollama is installed properly! 🥳

  ---

  🦙 Step 3: Pick a Model – TinyLLaMA

  Since LLaMA is a family of models (like llama2, llama3, codellama, etc.), let’s start simple and light.

  We’ll use TinyLLaMA – a lightweight model that works great on normal laptops 💻✨

  ---

  📥 Step 4: Pull the Model

  Now in your terminal, type:

    ollama pull tinyllama
  

  This will download the TinyLLaMA model to your system. It may take a few minutes depending on your internet speed.

  ---

  🚀 Step 5: Run the Model

  Once the download is done, just type:

    ollama run tinyllama
  

  Boom! You now have your own offline chatbot running right on your PC. 🎉

  You can start chatting with it directly in the terminal like:

    > Hello!
    < Hi there! How can I help you today?
  

  🧪 Step 6: Create a Virtual Environment and activate(Optional but Recommended!)

  Before we start writing our bot logic in Python, let’s first set up a virtual environment — a clean workspace for our bot project.

  This keeps your dependencies organized and avoids messing with your system Python 🧼🐍

  ---

  🪟 For Windows:

  Open Command Prompt (or PowerShell), then type:

    python -m venv chatbot
  
    powershell -ExecutionPolicy Bypass -File .\chatbot-env\Scripts\Activate.ps1
  

  You’ll know it’s activated when your terminal prompt shows something like:

    scssCopyEdit(chatbot-env) C:\Users\Sudharshini\Desktop>
  

  ---

  🐧 For Linux/macOS:

  Open your terminal and type:

    python3 -m venv chatbot-env
    source chatbot-env/bin/activate
  

  You’ll see:

    rubyCopyEdit(chatbot-env) sudharshini@ubuntu:~$
  

  🧪 Step 7: Install Python Packages for Your Bot

• 🔍 Installation

    pip install langchain langchain-ollama ollama
  

  🧠 Step 8: Write Your First Chatbot – main.py

  It’s time to bring your bot to life! 🪄
  Let’s create a Python file called main.py and add this code inside:

    from langchain_ollama import OllamaLLM
  
    def main():
        llm = OllamaLLM(model="tinyLlama")
  
        print("Welcome to the chatbot! Type 'exit' to quit.")
        while True:
            question = input("You: ").strip()
            if question.lower() == "exit":
                print("Goodbye!")
                break
            if question == "":
                continue
            print("Bot is thinking...")
            try:
                answer = llm.invoke(question)
                print(f"Bot: {answer}")
            except Exception as e:
                print(f"Error: {e}")
  
    if __name__ == "__main__":
        main()
  

  ---

  📝 What This Code Does

  • Creates a chatbot using your local TinyLLaMA model

  • Takes user input through the terminal

  • Sends it to the model and displays the response

  • Runs in a loop until you type exit

    ▶️ Run the Bot (Inside Virtual Environment)

    In your terminal, make sure you're inside your virtual environment, then run:

      python main.py
    

    You should see:

      Welcome to the chatbot! Type 'exit' to quit.
      You:
    

🌐 What’s Next?

Now that you've built a fully working chatbot in your terminal using LangChain + Ollama + TinyLLaMA...

💡 Why not make it talk through a website?

In the next blog, we’ll explore:

✅ How to create a Django web app
✅ How to integrate your local chatbot with a web frontend
✅ Make it interactive with a chat interface
✅ Still… no API keys! 😌

Stay tuned for Part 2: Integrate Your AI Bot with Django 🎀
From command line to cute website – we're just getting started 💻🌐💕