How to Install Ollama on Your System and Run Your AI Locally
Ollama is a powerful tool for running and managing AI models locally on your machine. Whether you're a developer, researcher, or AI enthusiast, installing Ollama is a straightforward process. In this guide, I'll walk you through the installation steps.
Visit the Ollama Website
Go to https://ollama.com and click on download—or if you're as lazy as I am, just click here: DOWNLOAD PAGE.
Download the version for your system. If, by any chance, you get a 403 error, don’t worry—just use a VPN and connect to a U.S. location before trying again. I personally use ExpressVPN, but any other provider should work fine.
Once installed, choose your model from the Models List. You can download a model by running:
ollama run YOURMODEL
You can list your installed models with:
ollama list
Models I’ve Used
| Model | Size |
|---|---|
| deepseek-r1 | 4.9 GB |
| Dodsworth | 2.0 GB |
| llama3.2 | 2.0 GB |
| deepseek-r1 | 9.0 GB |
| codellama | 7.4 GB |
| gemma2 | 5.4 GB |
| llama3.2-vision | 7.9 GB |
| phi4 | 9.1 GB |
Important: You should have at least 8 GB of RAM available to run the 7B models, 16 GB for the 13B models, and 32 GB for the 33B models—so choose accordingly.
Make sure to check out their official documentation DOCS for more details.
How to Customize a Model
A simple way to customize a model is by creating a model file:
code Modelfile
Then, add something like this:
FROM llama3.2
PARAMETER temperature 1
SYSTEM """
You are Alfred, the butler from Batman.
"""
To create the model:
ollama create alfred -f ./Modelfile
Now, if you run:
ollama list
you should see alfred (or whatever name you chose) in the list.
Run your model with:
ollama run alfred
You'll get something like this:
ollama run alfred
>>> Who are you?
Good day, I am Alfred Pennyworth, the loyal and dedicated butler to the Wayne family, and currently, the guardian of Master Bruce Wayne, also known as the Caped Crusader, Batman...
There are many other ways to train your AI, but this is probably the simplest.
LM Studio, OpenWebUI
At this point, you might be thinking, Gee, this is cool and all, but what if I don’t want to interact with my AI through the terminal? What if you want a GUI but don’t want to create one or work with APIs? Well, some brilliant folks have built tools that provide exactly that.
If You Want a GUI for Your Ollama Models
You can install OpenWebUI. Here’s how:
docker run -d -p 3000:8080 --add-host=host.docker.internal:host-gateway -v open-webui:/app/backend/data --name open-webui --restart always ghcr.io/open-webui/open-webui:main
this lets you interact with your Ollama models through a GUI. Be sure to check their documentation DOCS
If You Want a Standalone AI App on Your PC
At the time of writing, LM Studio does not support importing models from Ollama, so you'll need to download models separately.
To install LM Studio, visit LM Studio and download it for your OS. Click on Search Models and download the one you like. LM Studio allows you to specify a directory, which is a nice feature.
I hope you enjoyed this guide! Be sure to check out their websites and official documentation for more specifics this was just a quick intro.