Ollama Tutorial for Beginners — From Zero to Chatting with AI

Ollama is the fastest way to start running AI models on your own computer. This tutorial goes beyond basic installation — you'll learn how to have better conversations, use system prompts, switch models, and use the API.

Prerequisites

• 8 GB RAM minimum (16 GB recommended)
• macOS, Windows, or Linux
• Basic terminal familiarity

Install Ollama

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

# Windows — download the installer from https://ollama.com

Verify the installation:

ollama --version

Run Your First Model

ollama run llama3.2

Ollama downloads the model (about 2 GB) and starts an interactive chat session. Type your message and press Enter. The AI responds directly in your terminal.

To exit the chat, type /bye or press Ctrl+D.

Conversation Tips

Be Specific

Bad:

Help me with code

Good:

Write a Python function that reads a CSV file and returns the
rows where the "price" column is greater than 100. Include error
handling for missing files.

Ask for Formats

List the top 5 benefits of running AI locally.
Format as a numbered list with one sentence each.

Iterate

Now make it more concise.

Rewrite that for a non-technical audience.

Ollama remembers your conversation context within the same session.

System Prompts

System prompts set the AI's behavior for the entire conversation. They're powerful for customizing output.

Create a file called Modelfile:

cat > Modelfile << 'EOF'
FROM llama3.2

SYSTEM """
You are a concise technical writer. Always respond in bullet points.
Keep answers under 100 words. Use simple language.
"""
EOF

ollama create tech-writer -f Modelfile
ollama run tech-writer

Now every response from this model follows your system prompt rules.

Switching Between Models

You don't need to stop Ollama to switch models. In a chat session:

>>> /byemodel qwen2.5:7b

Or start a new session with a different model:

ollama run qwen2.5:7b

Popular models to try:

See the full list: ollama list (installed models) or browse at ollama.com/library.

Managing Models

# List installed models
ollama list

# Pull a model without running it
ollama pull llama3.1

# Delete a model to free space
ollama rm llama3.2

# Get info about a model
ollama show llama3.1

Using the API

Ollama runs an API server automatically. This lets you build applications that use local AI.

Start the server:

ollama serve

Make a request:

curl http://localhost:11434/api/generate -d '{
  "model": "llama3.2",
  "prompt": "Explain quantum computing in one paragraph",
  "stream": false
}'

Chat-style API:

curl http://localhost:11434/api/chat -d '{
  "model": "llama3.2",
  "messages": [
    {"role": "system", "content": "You are a helpful assistant."},
    {"role": "user", "content": "What is local AI?"}
  ],
  "stream": false
}'

The API is OpenAI-compatible, so most OpenAI client libraries work with Ollama by changing the base URL.

Adding a Web Interface

Prefer a graphical interface? Open WebUI adds a ChatGPT-like web interface on top of Ollama:

docker run -d -p 3000:8080 \
  --add-host=host.docker.internal:host-gateway \
  -v open-webui:/app/backend/data \
  --name open-webui \
  ghcr.io/open-webui/open-webui:main

Then open http://localhost:3000 in your browser.

Performance Tips

• Close other apps to free RAM for the model
• Use smaller models (like Llama 3.2) for quick tasks
• Apple M-series Macs get the best performance with Metal acceleration
• NVIDIA GPUs are auto-detected and used for acceleration
• First response is slower — the model loads into memory, then subsequent responses are fast

Summary

You now know how to:

• Install and run Ollama
• Have effective conversations with AI models
• Create custom models with system prompts
• Switch between models for different tasks
• Use the API in your own applications

Next Steps

• Best Models for 8GB RAM — detailed model recommendations
• Ollama vs LM Studio — compare with a GUI alternative
• Deploy Ollama on Runpod — run bigger models on cloud GPU