What you’re learning
Ollama runs the model and provides a local API. Open WebUI is a separate web app that talks to that API and gives you conversations, chat history, and a familiar interface. They are two pieces that need to be able to reach each other.
Work through it
- Before you start. Finish Lesson 01 and confirm `ollama run gemma2:2b` works. Install Docker Desktop (Windows or macOS) or Docker Engine (Linux) from its official source, then make sure the Docker service is running.
- Start Open WebUI. The official quick start offers a container that bundles Open WebUI and Ollama. This is the simplest isolated first try; it keeps its model and app data in Docker volumes. If you already have Ollama models installed outside Docker, follow Open WebUI’s “connect to Ollama” instructions instead so you do not download duplicate model files.
- Open the interface. Visit `http://localhost:3000` in your browser. Complete the first-run account setup on your own machine, then select Gemma 2 and send a short test prompt.
- Notice what changed. Try the same prompt in the terminal and the browser. The model is still Ollama; Open WebUI changes how you interact with it. Record any differences in response and setup.
Example command / prompt
docker run -d -p 127.0.0.1:3000:8080 -v ollama:/root/.ollama -v open-webui:/app/backend/data --name open-webui ghcr.io/open-webui/open-webui:ollama
Keep this in mind. This is the documented CPU-only bundled setup. If you choose an alternative setup, use the current official Open WebUI quick start. Keep the service bound to your own computer while learning; do not expose it to the public internet or add cloud API keys unless you intend to use those providers.
Try thisAsk Gemma the same question from terminal and web UI. Compare convenience, conversation history, and whether each response follows the same formatting constraints.