🦙Building a Local ChatGPT-like Environment with Ollama and Open WebUI on Ubuntu🦙

---

Introduction

This article serves as a personal memo on how to set up a local ChatGPT-like conversational environment using Ollama and Open WebUI on an Ubuntu system.
Ollama is a tool that allows you to run various LLMs (Large Language Models) locally, and Open WebUI provides a web-based interface to interact with Ollama.

Prerequisites

The following setup was verified on:

Although I mention "Ubuntu," I'm actually running it on a low-end Mac mini 2018 with Ubuntu installed—so the specs are quite modest.

lsb_release -d
Description:    Ubuntu 24.04 LTS

ollama --version
ollama version is 0.6.8

Setup Steps

1. Install Ollama

First, install Ollama using the following command:

curl -L https://ollama.ai/install.sh | bash

After installation, check if it works:

ollama --version

You can find models at ollama.com, and download or run them with commands like:

ollama pull <model_name>
ollama run <model_name>

Example:

ollama pull phi4
ollama run phi4

ollama pull gemma3
ollama run gemma3

2. Run Open WebUI

To start Open WebUI, run the following Docker command:

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

2.5. If Docker Cannot Connect to Ollama

In my case, Open WebUI inside Docker couldn’t connect to Ollama directly, so I tested connectivity using curl from inside the container.

1. Check if the container can reach Ollama:

docker run --rm curlimages/curl http://172.17.0.1:11434

2. If not reachable, modify Ollama settings to allow external connections.

Edit /etc/systemd/system/ollama.service and add the following:

Environment="OLLAMA_HOST=0.0.0.0"

You can also run the server directly like this for testing:

OLLAMA_HOST=0.0.0.0 ollama serve

Then reload and restart the service:

sudo systemctl daemon-reload
sudo systemctl restart ollama

Note: Setting OLLAMA_HOST=0.0.0.0 exposes Ollama to all external connections, which is suitable for debugging but not recommended for production. Please secure it appropriately.

3. Using Open WebUI

Once Open WebUI is running, open your browser and navigate to http://localhost:3000.
You’ll be able to select models already downloaded via Ollama and start chatting through the browser interface.

Conclusion & Thoughts

In this article, I explained how to install Ollama and Open WebUI on Ubuntu to create a local ChatGPT-like experience.
Since I ran this on a relatively underpowered machine, performance was predictably modest—the responses were slow and the models weren't particularly "smart."

I’ve started to explore the Agent mode feature, but so far I wouldn’t say it’s ready for serious use.
Still, I'm curious about potential fun use cases and plan to experiment more.

Note: This article is an English translation of my original post published in Japanese on Zenn: https://zenn.dev/ara_ta3/articles/ollama-openwebui-getting-started