Run Your Own AI on Your Computer: A Step-by-Step Guide to Ollama and Open-Source LLMs

Learn how to run powerful open-source AI models like DeepSeek, Qwen, and LLaMA completely offline on your own computer using Ollama. Step-by-step installation and hardware guide for local AI.

img 5416

“A clean desktop setup featuring a laptop running Ollama for local AI, accompanied by open-source guides for models like DeepSeek, Qwen, and LLaMA, illustrating private and secure computing on local hardware.” 

Ready to take full control in the AI revolution? The most practical way to say goodbye to cloud subscription fees, data privacy concerns, or the requirement of an active internet connection is running Local AI.

In this guide, we walk you through setting up Ollama—the most popular open-source tool for local AI—so you can run models like DeepSeek, Qwen, and LLaMA step-by-step right on your own machine.

1. Hardware Requirements: Is Your Computer Ready?

Open-source models directly utilize system resources, especially RAM and VRAM. Here are the general requirements based on model size:

  • Entry-Level (8GB – 16GB RAM / VRAM): Comfortably runs 7B – 8B parameter models (e.g., qwen3:8b, deepseek-r1:8b).
  • Mid-Range (32GB RAM / VRAM): Ideal for powerful 14B – 32B parameter models (e.g., deepseek-r1:14b, qwen3.8:27b).
  • High-End (64GB+ Unified Memory / 24GB+ GPU VRAM): Capable of running massive 70B+ models locally.

Note: Apple Silicon M-Series (M1/M2/M3/M4) Macs perform exceptionally well at running local AI thanks to their unified memory architecture.

2. Step 1: Installing Ollama

Ollama is a free, open-source tool operating on a container-like architecture that allows you to download, manage, and host local AI models as an API server.

macOS

You can install it via Homebrew in Terminal or directly using the official installer:

brew install ollama

(Alternatively, download the .dmg file from ollama.com.)

Windows

  1. Go to ollama.com/download.
  2. Download OllamaSetup.exe and complete the setup wizard.

Linux

Run the single-line installation command in your terminal:

curl -fsSL [https://ollama.com/install.sh](https://ollama.com/install.sh) | sh

Once installed, verify the installation by running this command in Terminal or Command Prompt:

ollama --version

3. Step 2: Downloading and Running Your First Model

Running a model with Ollama requires just a single terminal command. The ollama run command automatically downloads the model if it is not already present on your machine, then launches the chat interface.

Popular Model Commands:

  • DeepSeek-R1 (Reasoning & Logic Focused): ollama run deepseek-r1:8b (For more powerful hardware, opt for deepseek-r1:14b or 32b.)
  • Qwen (Coding, Multilingual & General Capabilities): ollama run qwen3:8b
  • Meta LLaMA: ollama run llama3.2

After running the command, the download will finish and a >>> prompt will open in your terminal. Type your question and chat directly! To exit the session, type /bye.

4. Step 3: Adding a User-Friendly Graphical Interface (GUI)

If you prefer a sleek, ChatGPT-like interface instead of terminal interactions, two excellent free options are available:

  1. Chatbox AI: A desktop application where you set Ollama (http://localhost:11434) as your API provider in the settings to chat with your local models visually.
  2. Open WebUI: A feature-rich, customizable ChatGPT clone that runs via Docker.

Useful Ollama Commands Quick Sheet

  • List installed models: ollama list
  • Download a model only (without running): ollama pull model_name
  • Remove a model from your system: ollama rm model_name
  • Check currently running models: ollama ps

Your data remains completely safe on your local drive, giving you a private AI assistant that works even when offline!