- Prerequisite
- Install the Official Ollama Python Package
- Verify That Ollama Is Running
- Your First Python Script Using Ollama
- How the Python Client Works
- Example: Build a Local Text Summarizer
- Example: Build a Simple Coding Helper
- Load Text From a File and Summarize It
- Common Errors & Fixes
- Mini Project: Create Your First Local Chatbot
- FAQ
- Conclusion
- References
In earlier articles, we installed Ollama and learned how to run models from the terminal. Now it’s time to take the next step using Ollama with Python. This is where things get really powerful. By connecting Python with Ollama, you can build:
- Local chatbots
- Document summarizers
- AI tools without internet
- Offline coding assistants
- Mini apps for personal use
And the best part?
No API keys, no billing, no rate limits. Everything runs locally.
Let’s begin.
Prerequisite
- pip must be installed on your machine
Install the Official Ollama Python Package
Open your terminal and run:
pip install ollama
This installs the official Ollama Python client. This works on all the operating systems
- Windows
- macOS
- Linux
Verify That Ollama Is Running
Before using Python, make sure Ollama is active.
ollama serve
Or check with:
ollama list
If a list of models appears, you’re ready.
Your First Python Script Using Ollama
Create a new file having name hello_ollama.py
from ollama import Client
client = Client()
response = client.chat(model='llama3.1', messages=[
{"role": "user", "content": "Hello! How are you?"}
])
print(response['message']['content'])
Run it:
python hello_ollama.py
You will see a friendly reply just like ChatGPT, but offline.
How the Python Client Works
Ollama uses a simple message structure:
messages = [
{"role": "user", "content": "question here"},
{"role": "assistant", "content": "previous answer"}
]
This allows conversation memory.
Example: Build a Local Text Summarizer
Create a file named summarizer.py
from ollama import Client
client = Client()
text = """
Artificial intelligence is transforming industries by enabling automation,
predictive analytics, and personalized solutions...
"""
prompt = f"Summarize this text in simple words:\n\n{text}"
response = client.chat(model='llama3.1', messages=[
{"role": "user", "content": prompt}
])
print("\nSummary:\n", response["message"]["content"])
Run:
python summarizer.py
Example: Build a Simple Coding Helper
from ollama import Client
client = Client()
question = "Write a Python function to reverse a string."
response = client.chat(model="phi3", messages=[
{"role": "user", "content": question}
])
print(response["message"]["content"])
Using Phi 3 gives better coding results.
Load Text From a File and Summarize It
from ollama import Client
client = Client()
with open("notes.txt", "r") as f:
content = f.read()
prompt = "Summarize the following text:\n" + content
response = client.chat(model="llama3.1", messages=[
{"role": "user", "content": prompt}
])
print(response["message"]["content"])
Perfect for:
- Meeting notes
- College notes
- Blog drafts
- Research documents
Common Errors & Fixes
Error: “Connection refused”
Ollama isn’t running.
ollama serve
Error: “Model not found”
ollama pull llama3.1
Error: Slow response
Use a smaller model:
ollama pull phi3
Error: Python can’t find package
pip install --upgrade pip
pip install ollama
Mini Project: Create Your First Local Chatbot
Create a new file named chatbot.py
from ollama import Client
client = Client()
print("Local AI Chatbot (Type 'exit' to quit)\n")
history = []
while True:
user_input = input("You: ")
if user_input.lower() == "exit":
break
history.append({"role": "user", "content": user_input})
response = client.chat(model="llama3.1", messages=history)
answer = response["message"]["content"]
print("Bot:", answer)
history.append({"role": "assistant", "content": answer})
Run:
python chatbot.py
You’ve now built your own local ChatGPT.
FAQ
Can I use multiple models?
Yes, simply set:model="mistral" or model="phi3"
Does Python work offline?
Yes, after you download the model.
Is it faster than cloud APIs?
For small & medium models — yes.
Conclusion
Using Ollama with Python opens the door to unlimited offline AI experiments. You can build chatbots, summarizers, coding helpers, and personal tools all without depending on cloud services.
Raj Ranjan