Unlock Self-LLM: Build a Chatbot in Under 1 Hour 2025

Ollama has made this step ridiculously simple. Open your terminal (Terminal on macOS/Linux, PowerShell or Command Prompt on Windows).

On macOS or Linux, run this single command:

curl -fsSL https://ollama.com/install.sh | sh

On Windows, just download the installer from the official Ollama website and run it.

Once Ollama is installed and running, you need to download a model. We'll use Meta's powerful and popular llama3 model (the 8-billion parameter version). In your terminal, run:

ollama pull llama3

This will download a few gigabytes of data, so it might take a few minutes depending on your internet speed. Go grab a coffee. When it's done, the model is ready and waiting on your machine.

Step 2: Set Up Your Python Environment

Now, let's create a space for our project. Create a new folder, name it something like my-chatbot, and open it in your favorite code editor.

We need two Python libraries: streamlit for the user interface and ollama to communicate with our local model. Install them using pip:

pip install streamlit ollama

That's it. Your environment is ready.

Step 3: Write the Chatbot Code

This is where the magic happens. Create a new file in your project folder named app.py. Paste the following code into it. Don't worry, we'll walk through what it all does.

import streamlit as st
import ollama

# Set the title of the Streamlit app
st.title("My Personal Llama 3 Chatbot llama3")

# Initialize the chat history in session state if it doesn't exist
if "messages" not in st.session_state:
    st.session_state.messages = []

# Display past messages from the chat history
for message in st.session_state.messages:
    with st.chat_message(message["role"]):
        st.markdown(message["content"])

# Handle user input
if prompt := st.chat_input("What can I help you with today?"):
    # Add the user's message to the chat history
    st.session_state.messages.append({"role": "user", "content": prompt})
    
    # Display the user's message in the chat interface
    with st.chat_message("user"):
        st.markdown(prompt)

    # Display the assistant's response in the chat interface
    with st.chat_message("assistant"):
        # Use a placeholder for the streaming response
        message_placeholder = st.empty()
        full_response = ""
        
        # Stream the response from the Ollama model
        stream = ollama.chat(
            model='llama3', # The model we downloaded
            messages=st.session_state.messages,
            stream=True,
        )
        
        # Append each chunk of the response to the full response
        for chunk in stream:
            full_response += chunk['message']['content']
            message_placeholder.markdown(full_response + "▌") # Add a typing cursor
        
        # Update the placeholder with the final, complete response
        message_placeholder.markdown(full_response)

    # Add the assistant's final response to the chat history
    st.session_state.messages.append({"role": "assistant", "content": full_response})

What's happening here?

• We import streamlit and ollama.
• st.session_state.messages is our memory. It's a list that stores the entire conversation so the model has context.
• We loop through the history to display past messages every time the app refreshes.
• st.chat_input() creates the text box at the bottom of the screen.
• When the user enters a prompt, we add it to our history and display it.
• The core is ollama.chat(). We send our entire message history to the llama3 model and set stream=True. This makes the model send back its response word by word, creating that cool live-typing effect.
• We display the streamed response and, once it's finished, save the complete answer to our history.

Step 4: Run Your Chatbot!

This is the moment of truth. Go back to your terminal, make sure you're in your project folder (my-chatbot), and run:

streamlit run app.py

Your web browser should automatically open a new tab. And there it is. Your very own, locally-hosted, private AI chatbot. Ask it anything!

You've Unlocked Self-LLM. What's Next?

Take a moment to appreciate what you just did. In under an hour, you've built a sophisticated piece of AI technology that, just a couple of years ago, was the exclusive domain of massive tech companies. This is a huge leap.

Where do you go from here? The possibilities are thrilling.

• Experiment with Models: Llama 3 is fantastic, but there are others! Try Microsoft's tiny but mighty phi3 or Google's gemma. Just run ollama pull <model_name> and change the model name in your app.py script.
• Give Your Bot a Personality: You can add a "system prompt" to guide the AI's behavior. Try adding a message like {"role": "system", "content": "You are a sarcastic pirate assistant."} to the start of your message history.
• Chat with Your Documents: The next big step is a technique called Retrieval-Augmented Generation (RAG). It allows your chatbot to read your own documents (PDFs, text files, etc.) and answer questions based on their content. It's the key to creating a truly personal knowledge assistant.

You've taken the first and most important step. You've moved from being a consumer of AI to a creator. You've built a foundation for privacy, ownership, and limitless exploration in the world of artificial intelligence. Now go build something amazing.